To improve the speed of the existing question and answering systems for emergency use, this paper uses knowledge graph technology to realize an emergency question and answering system adopting the interactive method of natural language. Based on vertical medical website, the system uses web crawler technology to build knowledge base, which is input into the Neo4j graphs for storage, presenting in the form of knowledge graphs. Based on rule matching method and string matching algorithm, the system clas- sifies and queries users’questions. The emergency question and answering system has application value in the construction of knowledge graph and question answering system in the field of emergency medicine.