Abstract: Accurate early warning and comprehensive management of mine disasters are essential for improving coal mine safety. To overcome the limitations and potential guidance bias of single early warning methods, an intelligent disaster management approach integrating big data analysis and numerical simulation is proposed. Taking the 3308 intelligent fully mechanized mining face of Shanxi Tiandi Wangpo Coal Industry Co., Ltd. as the engineering background, a big data analysis framework based on multi-source monitoring data was established. A three-dimensional numerical model was developed using the FLAC3D numerical simulation software to dynamically simulate the stress evolution of coal-rock masses and gas migration behavior during the mining process. By performing bidirectional verification between field monitoring data and simulation results, the proposed approach enables advanced warning and collaborative analysis of key hazards, such as gas and stress anomalies. Field applications demonstrate that this method can effectively identify the evolution trends of disaster risks, achieving an accuracy exceeding 95% in disaster attribute analysis and early warning.