Preventing coal and gas outbursts is critical to maintaining safe and effective mining and a reliable energy supply in China. In order to solve the problems of accident chain uncertainty reasoning and a priori information ambiguity in outburst risk analysis, this paper explored the macroscopic causation mechanism based on 80 cases of coal and gas outburst accidents, constructed three types of hazards causation models using topological networks, developed a fuzzy Bayesian risk assessment model. The probability of coal and gas outburst accidents was assessed using causal reasoning, diagnostic reasoning, sensitivity analysis, and key causal path analysis, and the causative mechanism was discovered. The case study showed that: the probability of outburst in the working face of this mine was 1.3%; at the time of the accident, the probabilities of the occurrence of the second and third hazard increased by 1050% and 725%, respectively; through the analysis of the key causal paths, the probability of outburst caused by path 1 (abnormal geological conditions and organizational and management deficiencies) and path 2 (abnormal geological conditions and inadequate gas extraction) rose from the usual condition by 462.8% and 569.2%, respectively. Finally, the model was validated using two outburst accidents as samples, and the results revealed that the probability of outburst for the two accidental coal mines was 12% and 10%, respectively, and the critical causal paths were largely consistent with the accident investigation reports. The assessment model presented in this study can help managers effectively control outburst accidents.