In response to the problem of weak correlation between related elements in current river pollution incidents, a river pollution knowledge graph is constructed and applied based on the advantages of knowledge representation in knowledge graphs. Firstly, the construction methods of the pattern layer and data layer of the knowledge graph are sorted out. Then, take the Mulanxi River Basin for example, the pattern layer of the river pollution knowledge graph is constructed in a top-down manner, the data layer is constructed using structured data and unstructured data extraction techniques, and the knowledge is stored and applied using the Neo4j graph database. Finally, rapid tracing of a pollution event in the Mulanxi River Basin is achieved using this knowledge graph. The experiment shows that the construction of this river pollution knowledge graph solves the problems of scattered data and weak correlation in water pollution events effectively, which can provide certain application value for rapid tracing of water pollution events.