After the Fukushima accident, the lessons involving in nuclear emergency management show that the emergency decision should enhance the capacity of analyzing and predicting severe accident. In order to improve this capability, the severe accident management training simulator (SAMTS) has been developed. This simulator has transplanted the self-developed integral severe accident code cosSA to the simulation platform to build the accident scenario, and coupled with computerized severe accident management guideline (CSAMG). The SAMTS provides several interventions to simulate different mitigation measures in SAMG, which help the operators handle to mitigate consequences and understanding the impact of mitigating actions on accident progression. This simulator could build the accident scene quickly to forecast and analysis to make central role of the information source for decision-making technical support in nuclear emergency management. The main purpose of this paper is to give a brief description of this simulator, including architecture, methodology, physical models of cosSA and a simulation case. Simulation results were compared with MELCOR (mainstream simulator calculation engine service) with the same initial and boundary conditions. Comparison results show that the calculation results of temperature, pressure and water level by SAMTS agree well with MELCOR. The good agreement proves the simulation capability of cosSA, which shows that cosSA could be applied into the severe accident simulator.