Given the characteristics of unconventional oil and gas reservoirs，such as low porosity，low permeability，and well-developed natural fractures，this study systematically reviews the theoretical and application progress of the hydraulic fracturing technology. The full text revolves four core domains：First，it reviews the development history of hydraulic fracturing numerical simulation methods and explores the control mechanisms of cross-layer behaviors of fractures in vertical wells and balanced multi-cluster propagation of fractures in horizontal wells. Second，it analyzes the characteristics of fracturing operation curves and the formation mechanisms of screen-out，while highlighting real-time intelligent early warning technologies based on artificial intelligence. Third，it summarizes the evaluation methods for stimulated reservoir volume （SRV） and post-fracturing effectiveness based on instantaneous shut-in pressure decline analysis （G-function），micro-seismic data，and dynamic production data. Finally，it compares the application advantages of traditional empirical production decline models，modern numerical simulation，and machine learning in predicting production of oil and gas wells with complex fracture networks. This study intends to provide a solid theoretical foundation and reference for exploring fracturing mechanisms in complex reservoirs，optimizing operational designs，and improving the development efficiency of unconventional oil and gas. It also points out future development directions，including refined simulation of multi-field coupling，engineering adaptability of intelligent early warning，and precise dynamic characterization of SRV.