Abstract: In view of the problems of difficult integration of multi-source heterogeneous data, lagging early warning response, low intelligentization level of decision-making and others in Menkeqing Coal Mine, an intelligent safety management and control system integrating Big Data platform (achieving unified data management), knowledge graph (completing semantic association and reasoning of security business), and electronic business modeling (driving intelligent early warning and closed-loop management and control) is proposed. The system adopts a unified three-layer architecture to construct an electronic model library covering effect risk, prevention and control early warning, and emergency response, graphical model design and zero code data adaptation mechanism are formed to achieve dynamic risk chain identification and intelligent decision-making. Graph reasoning and risk chain deduction algorithms are innovatively introduced to achieve intelligent identification, early warning, disposal, and closed-loop control of high-risk scenarios. The application results show that the early warning response time of the system is shortented by 60% (to within 3 minutes) compared to the traditional static rule system, the high-risk identification rate is improved by 27%, the manual data filling volume is reduced by 70%, the system online rate reaches 99.5%, significantly improving the safety level, management efficiency, and economic benefits of mines.