Control system is the key factor to control the smooth operation of large electromechanical equipment, especially with the deepening of new technology, unmanned control system has been widely concerned. If the unmanned control system fails, due to the lack of timely human intervention, it will have an important impact on the control performance in key environments. In recent years, there are many methods for fault diagnosis of electromechanical control system, but the traditional methods mostly rely on manual intervention and empirical retrieval, and the system fault data presents the disadvantages of small sample number and high proportion of unstructured data. Based on this, in order to effectively improve the accuracy of fault diagnosis, this paper adopts the knowledge graph method, realizes the knowledge extraction, entity relationship expression and fusion disambiguation of data through data migration and the method based on Secondary database, constructs the fault diagnosis architecture design of control system based on knowledge graph, realizes fault diagnosis and maintenance guidance, and improves the efficiency of fault diagnosis.