Abstract: In view of the problems of low fusion degree, insufficient self-analysis and optimization ability, and difficulty in tracing the causes of early warning of traditional coal and gas outburst early warning model, a dynamic early warning model of coal and gas outburst based on multi-source information fusion was established by combining association rule algorithm and evidence theory algorithm. In this paper, the method of setting association rule corresponding to qualitative and quantitative warning indicators was analyzed, and the optimization method of warning indicators and the basic confidence distribution rule were obtained through correlation analysis; the evidence theory identification framework for the early warning analysis of coal and gas outburst was established, the basic confidence distribution rules based on the correlation analysis results were determined, and the evidence synthesis method and the fusion decision method based on the probabilistic function were given; the basic confidence distribution function was used to trace the cause of early warning, and the dynamic updating method of the model was given. The test results show that the model can be used for outburst warning analysis to realize automatic screening of warning indicators, automatic fusion analysis and decision-making of multiple indicators, automatic tracing of warning causes and dynamic updating and optimization of the model, and it is feasible to use this method for coal and gas outburst early warning.