The existing electromagnetic big data management method is single and unable to make full use of electromagnetic data. The concept of partition management is introduced. The electromagnetic data is partitioned by geographic attributes using a clustering algorithm and managed by the graph database. The knowledge graph entities are transformed from the electromagnetic clustering results, and then the entity relationships are extracted to explore potential relationships among electromagnetic data. Aiming at the problem of the difficulty and low efficiency of electromagnetic interference source positioning, an improved Received Signal Strength Indication(RSSI) positioning algorithm based on knowledge graphs and big data real-time processing technology is proposed. The experiment simulates the process of interference source location under real electromagnetic data, and analyzes the performance of single-target interference source and multi-target interference source positioning. The experimental results show that the proposed method for locating electromagnetic interference sources based on the knowledge graph is more effective than the traditional RSSI locating method, and its error is smaller.