A power equipment quality integrated management system based on picture, event knowledge, network and information are established. A top-down domain knowledge graph based on graph database is constructed to meet the actual requirements of power grid equipment quality assessment based on the data storage and dissemination mechanism of graph database. Therefore, it is possible to store multi-source heterogeneous data in graph database effectively. A knowledge graph small sample training and fault prediction method combining the influence factors of equipment faults and historical operating data is proposed based on the gradient lifting decision tree. Through the actual operation test of the substation equipment operation and maintenance processing system, it is found that the system can better promote the substation equipment operation and maintenance method based on event knowledge graph.