Abstract: The development of simulation scenarios has such problems, as strong professionalism, high customization requirements and poor mobility, etc. Aiming at these problems, an intelligent generation method of simulation scenarios based on knowledge graph（KG）is designed and implemented. Firstly, the KG is constructed by linking multi-source simulation model entities with global features based on reinforcement learning（RL）. Based on this, an entity alignment algorithm suitable for simulation model matching and a script generation method based on document object model（DOM）are designed to realize intelligent mapping of simulation scenarios. Finally, the "multi-dimension projection" action is taken as an example, the experimental results show that the generated simulation scenarios can drive the simulation system to run according to the scheduled action sequence, and the method is effective.