In today’s process systems, operators must consider an overwhelming amount of information which is passed to them via automated systems, and make decisions very quickly. Since the decision-making in a time-critical situation is extremely complicated, the use of automated systems to aid decision making is highly recommended. This paper proposes a hybrid Bayesian network (HBN) to support process operators in hazardous situations. The proposed HBN includes three parts: an evidence preparation, a situational network, and risk estimation. The evidence preparation part provides soft evidence based on the online conditions and process monitoring system. The situational network is developed based on dynamic Bayesian networks to model the hazardous situations, and the risk estimation part calculates the risk level of every situation dynamically to show whether the risk level of situations is acceptable or not. The threefold HBN is explained through a case from U.S. Chemical Safety Board (CSB) investigation report. According to the CSB report, following an operator error at a paint manufacturing plant, the explosion and subsequent fire destroyed a facility, injured ten residents, and heavily damaged dozens of nearby homes and businesses. Finally a sensitivity analysis is presented to evaluate the proposed HBN.