The sudden subway disaster resulting from exceptionally heavy rainfall in Zhengzhou,Henan Province,on July 20,2021,marked one of the most significant casualties the city had experienced in recent years.This incident arose from a combination of objective factors,such as intense rainstorms and unexpected accidents,and subjective factors,including a lack of risk awareness and inadequate emergency response mechanisms.This paper begins by identifying key scenario elements within the evolution of the “Zhengzhou Rainstorm Subway Disaster Event” and constructing the scenario evolution process of the event through scenario analysis.Building upon this foundation,we create a scenario deduction model for the “Zhengzhou Rainstorm Subway Disaster Event” using Bayesian network.Expert scoring methods are employed to calculate the conditional probabilities of network nodes,and Netica software is used to compute the state probabilities of disaster scenarios nodes.Finally,we assess and deduce the probabilities of personnel fatalities and subway damage under varying levels of rainstorm disasters,different emergency response activities,and distinct surrounding environmental conditions.The objective is to extract insights,pinpoint vulnerabilities,and provide valuable references for risk prevention and response measures in anticipation of and protection against events similar to the “Zhengzhou Rainstorm Subway Disaster”.