Abstract: An ontology-based emergency response plan knowledge modeling is proposed, and a unified and standardized metro emergency response plan ontology knowledge database is constructed. By referring to MyCBR case reasoning tool, with the established knowledge database of metro operation accident cases, a case retrieval method based on semantic similarity is proposed to realize the efficient retrieval of metro emergency accident cases. Using Protégé software and MyCBR tool, the function of knowledge reasoning and case retrieval of emergency response plan is realized, which significantly improves the knowledge application efficiency.