ABSTRACT Quick and efficient decisions should be made when unconventional emergencies occur. This paper develops an emergency group decision-making mechanism for use in unconventional emergency responses. The scenario analysis approach provides analysis results of current situations and future trends of emergency events as important decision information. The paper also utilizes Dynamic Bayesian Networks theory to build a scenario-response model of emergency group decision-making. A case study of the “7·23” high-speed railway accident in China is conducted to test the effectiveness of the proposed model. The analysis results show that the scenario-response model contributes to initiative and efficiency in unconventional emergency responses. The results also suggest that current situations and possible trends of unconventional emergencies should be considered as decision support information in the emergency decision-making process. Emergency response managers can utilize the research model to learn experience from existing emergency events and develop decision-making mechanisms for emergencies in the future.