With the introduction of AI and big data technologies, the oil and gas exploration field has realized a paradigm shift in the drilling process from being experience-driven to data-driven. At the same time, geology-engineering integration services also have shifted from "drilling wells successfully" to "drilling wells quickly". However, the integration technology still remains at the optimization and improvement phase of the working mode and operation process. During drilling in complex formations, there are still technical problems such as limited penetration rate, low trajectory control accuracy, and lagging response to downhole risks, which restrict the further development of this field. AI technology provides a new path to break through the bottlenecks： Through dynamic Mechanical Specific Energy （MSE） modeling and real-time optimization, accurate mapping from formation characteristics to engineering parameters is achieved. With this as the core, an intelligent drilling optimization system driven by the three cores of "data, decision and execution" has been built to achieve full-process closed-loop optimization. It has been successfully applied in the on-site drilling construction process, verifying that this system can improve penetration rate, reduce non-production time, and effectively control the deviation of the wellbore trajectory. This study provides theoretical support for the development of the intelligent drilling optimization system, reveals the transformation of digital technology to the traditional drilling optimization working mode, and provides replicable solution at the engineering level.

With the deepening of oil and gas exploration and development, the traditional gas logging interpretation methods mainly based on charts are unable to meet the evaluation needs of the relationships between the complex reservoirs and oil, gas and water. In recent years, the rapid development of big data analysis and machine learning has provided a new direction for gas logging interpretation. Based on gas logging interpretation data from multiple wells, data preprocessing and feature engineering methods are applied to process the gas logging data, and five common machine learning models （KNN, NB, SVM, MLP, LightGBM） are used for modeling interpretation,application,comparison and analysis. The results show that the comprehensive performance of SVM and LightGBM models is good with an average prediction accuracy of over 87%, which can meet the actual interpretation needs. The machine learning model solves the problems of the insufficient flexibility of the traditional interpretation charts and the difficulty of extracting deep information due to the dimensional limitations, and significantly improves the interpretation accuracy, which has the practical value of promotion and application.

In the field of geologic exploration, density logging curve is of great significance in determining reservoir porosity, identifying gas reservoirs, judging lithology, dividing oil-water interfaces, identifying fluid types, and improving inversion accuracy. However, in practical operation, logs may be lost or distorted due to instrument failures, data transmission errors or external interference. To solve this problem, this paper proposes a density logging curve reconstruction method based on Transformer-LSTM fusion model. This method utilizes Transformer′s self-attention mechanism to effectively capture the long-distance dependency relation in log data, and combines with the recursive characteristics of Long Short-Term Memory （LSTM） to significantly improve the reconstruction accuracy. By preprocessing, model construction and training of log data from Y zone of Ordos Basin, the performance is compared with bidirectional gated recurrent unit （BiGRU）, deep neural networks （DNN）, multiple regression analysis （Logistic）, temporal convolutional network （TCN） and Transformer model. The results show that Transformer-LSTM fusion model performs well in density logging curve reconstruction, especially in terms of reconstruction accuracy and generalization ability. The experimental results verify that the model is capable of reconstructing high-precision density curve data, providing reliable support for geologic exploration.

To improve the precision of reservoir identification by mud logging in oil and gas exploration, a mud logging system based on infrared spectroscopy technology has been developed. The hardware of the system adopts a miniaturized and integrated design with a high-sensitivity infrared detector, realizes the comprehensive identification of gas spectral features through infrared spectroscopy scanning technology, and completes the rapid and automatic analysis of hydrocarbons in drilling fluid. The software of the system has functions such as remote data sharing, software mapping and data analysis, which is used to realize stratigraphic interpretation and evaluation. The field test shows that the infrared spectroscopy gas logging system can quickly find the show of gas and oil, and has obvious technical advantages in the evaluation of thin beds and fracture-type reservoirs. It also has the advantages of high precision, low error, rapid response and good repeatability. It can meet the needs of mud logging operations in different regions and effectively improve the accuracy of reservoir evaluation.

With the improvement of drilling technology,the drilling speed of drilling exploration has been greatly increased,especially in the offshore drilling platforms and shallow well drilling,the drilling time per meter is extremely small （generally 0.1-1.0 min/m）,while the delay time of the gas collection pipeline for mud logging samples is usually the range of 2-10 min. When a single pipe is connected or the pump is stopped for special reasons,the sand return depth in the mud logging acquisition software will stop updating. However,the gas corresponding to the sand return depth in the sample gas pipeline is still being analyzed by chromatograph. The result is that the gas logging data analyzed by the chromatograph cannot correspond to the lagged well depth,and the gas logging data is "lost". The amount of "lost" data is determined by the lag time of the gas pipeline and the drilling time. In order to ensure the complete collection of gas logging data,an automatic control device for mud logging sample gas collection has been developed. The device collects real-time drilling status through RS-232 serial communication,and uses a microcontroller control unit in the integrated circuit to control the electromagnetic valve on the automatic control device to switch the sample gas pipeline,achieving accurate analysis of the sample gas in the pipeline and avoiding the "loss" of gas logging data.

Pure hardware logic control circuit technologies such as time-delay relays used in early comprehensive mud logging unit control systems are no longer applicable. In order to meet the dangerous working environment of the explosion-proof zone 1 on site, a new comprehensive mud logging unit control system was developed and designed. The system is centered on highly reliable Programmable Logic Controller （PLC）, by fusing high-sensitivity sensor detection, high-resolution data sampling, multi-computer communication, power redundancy and other technologies, and uses an embedded industrial control configuration touch screen as a man-machine interactive interface and data storage terminal. The safety isolation barrier is used as a safety guarantee for signal transmission on the safe side and the dangerous side. All electrical control components are placed inside the flame-proof box, with various explosion-proof detection sensors supplemented outside. The control system can automatically identify the danger level of the environment, make intelligent decisions based on relevant standards, and promptly implement different levels of protection measures such as reminders, alarms, and power outages. The first batch of 12 comprehensive mud logging units using this control system have been successfully applied to the mud logging service project of Kuwait Petroleum Corporation. The field effect is good, protecting the safety of site personnel effectively and ensuring the reliable operation of site equipment.

Aiming at the technical problem that it is difficult to quantitatively analyze the contamination degree of wireline formation test samples in oil-based drilling fluid environment due to the miscibility of crude oil and drilling fluid, this paper proposes a comprehensive quantitative evaluation method based on conventional fluid physical comparative analysis and saturated hydrocarbon gas chromatography component comparative analysis. Through the establishment of the sample density analysis chart and the gas chromatographic analysis chart, the calculation software of formation crude oil drilling fluid contamination proportion was developed, and the quantitative evaluation of the effects of oil-based drilling fluid on the contamination degree of the wireline formation test samples was realized. The method has been applied to 14 wells in Weixinan Sag of Beibu Gulf Basin. The analysis results show that the maximum absolute errors are within 5% compared with the laboratory sample analysis results and oil testing productivity analysis data. The field application results show that the method has high analytical accuracy and practicability, and can realize the rapid and accurate evaluation of oil-based drilling fluid contamination rate in wireline formation test samples.

In the exploration of Pearl River Mouth Basin, a large amount of CO₂ gas was found, and its genesis is complex, covering organic origin, inorganic origin and mixed origin. The study on CO₂ formation mechanism in this area is of great significance to judge the source of oil and gas and the mode of accumulation. However, the CO₂ monitored during mud logging while drilling may originate from the added drilling fluid materials, which has a significant influence on CO₂ genesis study. Because CO₂ inherits the geochemical information of maternal materials during its formation, different ${{\delta }^{13}}{{C}_{C{{O}_{2}}}}$ can indicate different biological sources. Through the design of relevant experiments, the ${{\delta }^{13}}{{C}_{C{{O}_{2}}}}$ values of the drilling fluid materials are obtained. Compared with the ${{\delta }^{13}}{{C}_{C{{O}_{2}}}}$ values measured during mud logging while drilling process, the sources of CO₂ can be confirmed. The ${{\delta }^{13}}{{C}_{C{{O}_{2}}}}$ values in the drilling fluid materials from wells A and B are quite different from the ${{\delta }^{13}}{{C}_{C{{O}_{2}}}}$ values while drilling, which indicates that CO₂ originates from formation rather than the drilling fluid materials. According to this, it is confirmed that CO₂ in well A of Enping Sag is the organic origin of organic matter pyrolysis, and CO₂ in well B of Eastern Yangjiang Sag is the inorganic origin of mantle source. This study provides important geochemical basis for oil and gas exploration in Enping Sag and Eastern Yangjiang Sag.

Gulong shale mainly consists of large suites of clayey felsic shales with many pore types and complex pore structures,and its reservoir performance is the core in shale oil geological research,but conventional logging methods are difficult to effectively evaluate such micro-nano pores. To solve this problem, NMR technique is applied to geophysical logging to full-scale characterize the porosity and pore structure of shale reservoirs. By building the conversion relationship between NMR relaxation time （T₂） and rock pore diameter, the proportion of large pores in Gulong shale was characterized by the specific T₂ spectrum components to evaluate the shale reservoir performance quantitatively, systematically revealing the pore structure differences and distribution law of Gulong shale reservoirs. Based on the results of the reservoir performance evaluation,the target strata with large porosity and good pore structure were selected to optimize the horizontal well trajectory designs. The results show that with the changes of rock mineral compositions, there are differences in the vertical pore structures of Gulong shale reservoirs. Among them, the middle-lower part of Q₉ oil layer has the highest proportion of large pores and the best pore structures. The overall production of horizontal wells in the expanded test area,which is implemented by using this part as the core target window of horizontal wells,has been greatly improved. This technology will provide favourable support for the follow-up efficient development of Gulong shale oil.

With the deepening of oil and gas exploration, it has become an inevitable trend to develop from middle-shallow to deep layers. In particular, major breakthroughs have been made in the deep exploration of the Permian in Mahu Sag and the Carboniferous in Shawan Sag, Junggar Basin, and the deep exploration has become an important replacement field for oil field reserve increase and production increase. However, in the Permian and Carboniferous strata, there are diverse rock types, complex compositions, fine drill cuttings, leading to difficulty in quickly identifying lithology, which restrict the efficiency deep oil and gas exploration. Therefore, X-ray diffraction mineral logging and Gamma spectrometry logging while drilling technologies have been introduced for research on the deep and mainly developed Permian and Carboniferous strata. Combined with thin section identification, the lithology identification charts can be established by dividing blocks and layering positions through the rock-mineral index model and the optimization of typical minerals. This provides a new method for rapid identification of lithology and precise horizon determination, guaranteeing drilling safety and further improving mud logging technology system. At present, a total of 13 wells have been applied, with the coincidence rate of over 82% for lithology identification. In the field of deep oil and gas exploration and development, the application effect is obvious.

Regarding the identification of fluid properties in low and ultra-low permeability reservoirs in Liu-2 and Liu-3 Members in Weixinan Sag,Beibuwan Basin,conventional logging techniques and interpretation methods are severely limited under the contamination of oil-based drilling fluid,resulting in insufficient accuracy of fluid property discrimination and difficulties in real-time decision making on site. Using the gas characteristic spectrum discriminant method,original absorption rate spectrum discriminant method,and oil and gas index quantitative discriminant method of the hydrocarbon phase state evaluation technology,the fluid properties of the low-permeability reservoirs in this block are analyzed,and the extended applications of this technology in fracture identification and oil and gas sources are discussed. The results show that the hydrocarbon phase state evaluation technology has played a unique advantage in oil and gas classification and reservoir water content identification. For the identification of fluid properties in more than 20 wells of various well types and reservoirs in different formations in Weixinan Sag,the coincidence rate of the comprehensive interpretation is over 90%. This method effectively solves the problem of fluid property identification caused by the dual factors of oil-based drilling fluid pollution and low-permeability reservoirs in Weixinan Sag,provides reliable technical support for offshore oil and gas field development and has significant value for popularization and application.

In order to solve the problem of accurate identification of fluids with different properties in complex reservoirs in oil and gas exploration,based on the application data of infrared spectrum technology in the Bohai Bay Basin,five sensitive wavelengths were optimized and four sets of key parameter combinations were screened by analyzing the spectral absorption efficiency curve characteristics of fluids with different properties. Through constructing the multiparameter fusion charts,the rapid identification and accurate evaluation of oil layers,high-GOR oil layers,gas layers,and oil-bearing water layers have been achieved. Validated by data of on-site testing while drilling in 6 wells,the interpretation coincidence rate of infrared spectrum technology reached 92.3%. This method not only improves the qualitative analysis ability of reservoir fluid,but also provides more efficient technical support for hydrocarbon reservoir exploration and development,and has important engineering application value.

Aiming at the problem of the low coincidence rate of the interpretation of the reservoir fluid properties in gas exploration wells in Ordos Basin, the key influencing factors of NMR logging technology were systematically analyzed. Based on core experimental data, a quantitative evaluation model centered on NMR porosity and the dynamic ratio of the initial state movable water saturation to the initial state irreducible water saturation was reconstructed, breaking through the excessive reliance on data quality by the traditional intuitive identification method of spectra. Field application shows that the model significantly improves the accuracy of reservoir water content discrimination by standardizing data acquisition process and establishing reservoir gradation and classification standards, and the interpretation coincidence rate increases from 79.23% to 85.32%. The research results confirm that the dynamic ratio parameter model can effectively characterize the fluid distribution characteristics of low-porosity and low-permeability reservoirs, thereby providing reliable technical support for the optimization of well completion test intervals and the establishment of reservoir stimulation schemes under complex geologic conditions, and has promotion value for the exploration and development of similar hydrocarbon reservoirs.

The identification of fluid properties in low-contrast reservoirs is a technical bottleneck in the field of oil-gas exploration and development, its core characteristics are manifested in the weak differences in resistivity between oil layers and water layers, which are difficult to effectively distinguish on conventional log response. In response to this problem, taking the shallow low-resistivity oil layers in Hanjiang Formation of Enping Sag and the Paleogene high-resistivity water layers in Lufeng Sag, Pearl River Basin as examples, a "3-axis linkage" mud logging evaluation technology system is constructed: gas logging identifies hydrocarbon anomalies, real-time fluid logging quantitatively characterizes hydrocarbon abundance and geochemical logging identifies the phase states of the fluids （oil or water layers）. On this basis, the multiparameter quantitative interpretation charts have been established, which effectively solve the evaluation difficulties in the reservoirs of thin interbedded layers and low-resistivity oil layers, and the accuracy of regional fluid property identification has been increased to more than 85%. In practice, well A 3H in Enping Sag has obtained a high-yield oil flow of 823 t/d, and in Lufeng Sag, the interfaces of high-resistivity water layers have been successfully identified, which make the identification accuracy of high-resistivity water layers with 55% of the traditional interpretation based on logging resistivity data increase to 82%, and provide a reliable technical support for the high efficiency development of the complex hydrocarbon reservoirs in South China Sea.

The exploration targets of Bohai Oilfield are shifting towards oil reservoir zones with complex geologic conditions, resulting in a significant increase in the difficulty of oil-water layer interpretation and evaluation. The traditional comprehensive evaluation method combining mud logging and well logging technologies restricts the operation efficiency to a certain extent due to the hysteresis of well logging technology. Therefore, a quantitative calculation model of oil saturation based on geochemical logging was established. This model comprehensively considers the distribution laws of oil-gas-water three-phase flow in reservoir pores and the influence of petrophysical properties （crude oil density, reservoir porosity, matrix density, etc.） on oil saturation, providing a new angle of view and tool for accurate assessment of reservoir fluid properties and realizing the real-time evaluation of reservoir fluid properties. Verified by the data of the wells newly drilled in X block of Bohai Oilfield, the accuracy rate of model interpretation was 86.67%.The oil saturation calculation model can provide strong support for oil-gas exploration and development in Bohai Oilfield.

Pai 612 block of Chunfeng Oilfield is the first block to develop super-extra viscous crude oil reservoirs in western Junggar Basin. Currently,multi-cycle steam injection exacerbates the heterogeneity of the reservoirs,and there are development problems such as frequent inter-well steam channeling and uneven reserve production. In order to deepen the understanding of reservoir heterogeneity,it is urgent to carry out differential lithofacies study and geologic modeling of reservoirs. The reservoirs of Pai 612 block are mainly divided into high-quality sandstone facies and calcareous sandstone facies. On this basis,the seismic data analysis and horizontal well stratigraphic correlation were used to improve the accuracy of structural control,and the structural model was established. Combined with the lithofacies study and the analysis of core data and log information,the linear correspondence between the lithofacies and the gamma curve of horizontal wells was determined,and the lithofacies data of the horizontal wells was quantitatively calculated to establish a differential lithofacies model for the block. The vertical well attribute model was established by regressing interval transit time and physical properties based on the analyzing and rock sample testing data of the cored wells, and the horizontal well attribute model was established based on the correlation between shale content and physical properties. Finally,by comparing the static and dynamic characteristics of different lithofacies reservoirs,as well as analyzing the production capacity,the net to gross ratio of different lithofacies was determined,and the production of calcareous sandstone facies accounted for 66.7% of high-quality sandstone facies. The different lithofacies differentiation net to gross ratio model of the final completion lays the foundation for improving remaining oil recovery efficiency in the later stage.

The Middle Indus Basin is located in the central part of the Indus Basin, which is a foreland basin developed from the passive continental margin. Previous research suggests that the factors controlling the formation of tight gas reservoirs in foreland basins are fault transport of oil and gas and fracture transformation of reservoirs, with oil and gas enriched in fracture developed zones near the faults. However, the drilling results in the local area are different from the previous research results. Not all faults in the area are enriched with oil and gas. Most of the wells successfully drilled in this area are distributed along NS-trending faults, and exploration near the NW-trending faults has not been broken through. Therefore, it is urgent to conduct more in-depth research on the differences in fault-controlled reservoirs, and explore tight gas reservoir formation models suitable for this area. Based on the understanding of regional geology and fine interpretation of seismic data, the characteristics and development stages of faults in M area have been summarized, with a focus on the structural evolution processes and stress conditions that give rise to these fault characteristics. This paper analyzes the sealing ability of these faults, and finally reveals fault controlling regulation on differential hydrocarbon accumulation in this area. The research provides a theoretical foundation for tight gas exploration in the Middle Indus Basin. The results indicate that NW-trending faults suffered compressive stress in Early Cretaceous epoch, leading to significant shale smearing and a certain sealing capability. The nearly NS-trending faults functioned as the main migration pathways for gas, controlling the hydrocarbon migration and accumulation. The tight sandstone reservoirs on both sides of these faults have been transformed by faults and fractures, and have great accumulation potential, which point out the way for the next exploration and development of the gas field.

The source direction, sedimentary facies evolution history and distribution law of oil-bearing series of Shahejie Formation in Wen'an slope, Jizhong Depression, are still unclear. Based on core, well logging, seismic and other geologic information, the provenance system and sedimentary facies characteristics of Shahejie Formation were discussed in depth through paleogeomorphological restoration method and heavy mineral Q type clustering procedure. The results indicate that the sedimentation of Shahejie Formation in Wen′an slope has multi-source characteristics, and the provenance mainly comes from the northwest of Niutuozhen uplift and the southeastern Wen′an slope. Two major source areas developed multiple transport pathways. Sediments have developed several graded sand bodies along the slope topography. Influenced by the steep western and wide-flat eastern paleogeomorphological pattern, Shahejie Formation mainly developed three sedimentary facies during its sedimentary period： fan delta, braided river delta and lake. During Member 3 of Shahejie Formation sedimentary stage, five sedimentary subfacies were mainly developed in the study area： fan delta plain and front, braided river delta plain and front and shore-shallow lacus. Multiple delta lobes were formed along the subaqueous distributary channels. During Member 2 of Shahejie Formation sedimentary stage, the northwest of the study area mainly developed shore-shallow lacustrine mud microfacies, while the southeast mainly developed two sedimentary subfacies of braided river delta plain and front. During Member 1 of Shahejie Formation sedimentary stage, the northwest of the study area mainly developed two sedimentary subfacies of fan delta plain and front. The southeast mainly developed braided river delta plain, with a small amount of braided river delta front sedimentary subfacies. The study results can provide theoretical basis and guidance for the further refined exploration and development of hydrocarbon resources in Shahejie Formation of Wen′an slope in Jizhong Depression.

Liuheying-Daxinzhuang structural belts are second-level oil-bearing structural belts located at the southwestern end of the western zone in Langgu Sag, with low degree of exploration. Due to its complex structure and sparse survey grid, the hydrocarbon accumulation regularities remain unclear. Based on the data of well-seismic combination, actually drilled geology and analytic test, the hydrocarbon accumulation conditions and exploration potential are deeply analyzed. The results are obtained in four aspects. First, The Es₃^l dark mudstone is the dominant source rock beds and has good oil source conditions with a large thickness of 600-1 600 m, high organic matter abundance of TOC 0.64%-1.03%, and main kerogen type Ⅱ₂-Ⅲ. Second, As a whole, Es₃ reservoirs are dominated by medium-high porosity and low permeability, with the best physical properties in Es₃^m of Daxinzhuang arched structural belt, and the reservoir-cap rock assemblage in Es₃ is superior. Third, The structural belts are clamped between the two large oil-generating troughs of western Gu′an and southern Liuheying, the oil and gas enrichment is jointly controlled by the nose-like structural trap, the turbidite sand body distribution, and the dual effects of oil source fault transport and later tectonic uplift, forming the down generated up stored and side generated lateral stored reservoir-forming patterns. Fourth, two reservoir-forming patterns of ridge-type and toothbrush-shaped structures were constructed to reveal the banded enrichment regularity of oil and gas along the structural high parts and faults. G 8 trap and No.2 trap in northern L 1 were selected as the most favorable targets from 49 traps, with the total quantity of oil resources of over 500×10⁴ t and natural gas of nearly 40×10⁸ m³ predicted. The results provide theoretical guidance and technical support for the subsequent research and exploration of oil-bearing traps in the complex structural zones of Langgu Sag.

A large number of glutenite segments are developed in the northern actic region of Dongying Sag, which is a favorable place for hydrocarbon enrichment and plays an important role in oil and gas exploration. However, in recent years, no effective hydrocarbon shows have been found in many wells in Shengtuo area during the drilling process of glutenite body under the lower section of Es₄. To explain the cause of this anomaly, based on the theory of hydrocarbon accumulation, combined with tectonic evolution, sedimentary history and logging data, this paper analyzes the reservoir-forming control conditions of source rocks, reservoirs, caprocks and traps in the lower section of Es₄ in Tuoshen 3 block. It is clarified that the self-sourced hydrocarbon reservoirs are mainly developed in the block. Compared with the source rocks of the low position in the upper section of Es₄, the source rocks developed under the salt-gypsum formations in the lower section of Es₄ with poor oil-bearing property and small oil-generating potential, so it is difficult to provide oil source support. Under the depositional setting of deep and semi-deep lacustrine, the thick-bedded salt-gypsum formations deposited on the top of the glutenite body constitute a high quality caprocks, and the isolated and discontinuous reservoir bodies developed under it have poor physical properties. The oil and gas are mainly migrated to the traps for preservation through the transporting system composed of faults, sand bodies and unconformity surfaces, and the structural-lithologic trap and sedimentary evolution model jointly control the hydrocarbon accumulation in the block. The study provides an important reference for the exploration,identification, and the later productivity construction of glutenite body in the lower section of Es₄, Tuoshen 3 block.

In order to fully understand the hydrocarbon accumulation process in Nanpu No.3 structural belt, and improve the effectiveness of oil and gas exploration, based on the data of geology and geochemistry, and using the technique of petroliferous basin numerical simulation, this paper reconstructs hydrocarbon-generating and hydrocarbon-expulsing histories, and discusses the matching relation of hydrocarbon accumulation on the basis of the systematic analysis of the source rock. The results show that there are three sets of source rock,including Es₃, Es₁ and Ed₃ in the study area,and the main types of organic matter are Ⅱ with high organic matter abundance. The Es₃ source rock have high hydrocarbon-generating and hydrocarbon-expulsing intensity at the stage of maturity and high maturity, and the mass process of hydrocarbon generation and expulsion happened from the end of Dongying sedimentation to the middle of Minghuazhen sedimentation. The Es₁ and Ed₃ source rocks have medium hydrocarbon-generating and hydrocarbon-expulsing intensity at the stage of maturity, and the main process of hydrocarbon generation and expulsion happened from the middle Minghuazhen sedimentation to present. The trap forming epoch matches well with generation, expulsion, migration and accumulation of hydrocarbon. The deep and middle-deep oil and gas exploration in the study area should take the source rocks of the Es₃ as the oil source, and the middle-shallow oil and gas exploration should take the source rocks of Es₁+Ed₃ as the oil source. The favorable exploration objects include structural and lithologic traps that are adjacent to Caofeidian sub-sag and have faults vertically connecting the source rocks and reservoirs.

In order to explore the dynamic variation law and influencing factors of the pressure field in the development process of low-permeability reservoirs, this paper focuses on Chang 2^1-3 oil reservoir set in wellblock W of Ansai Oilfield as object, and analyzes the spatio-temporal evolution characteristics of the formation pressure field in 2019 and 2023 by comprehensively using static pressure test, pressure drawdown test, and pressure build-up test methods. Combined with parameters such as geologic structure, sand body distribution and injection to production ratio, the main control factors and action mechanism of the pressure field dynamic variation are revealed. The following results are obtained： There is significant temporal heterogeneity in the pressure field, with flow coefficient differences reaching up to 22.5 times and notable pressure build-up in the local high-permeability zones; In spatial distribution, the pressure distribution has transitioned from the uniformity to the differentiation between high pressure and low pressure. The formation pressure in the eastern area increased by 11.04 %. The formation pressure decreases in the western and middle areas were 15.45 % and 21.56 % respectively. The lows in geologic structure, sand body thickness, and injection to production ratio were identified as main control factors of the pressure field dynamic variation, with a strong positive correlation observed between injection to production ratio and percentage increase in pressure （R² of 0.972 2）. The study offers a theoretical basis for the optimization of development adjustment scheme in low-permeability reservoirs.