The research on the full process tracking scheme of materials for power grid infrastructure is aimed at improving the reliability and safety of power grid operation, in response to the growing demand for electricity and the challenges of material management. This study aims to use knowledge graph technology to establish a comprehensive and dynamic material information network, achieving full lifecycle tracking and management of power grid materials from procurement to scrapping. In this paper, by integrating multi-modal information with attention mechanism, triplet knowledge graph is extended to temporal knowledge graph, and a corpus and dictionary are constructed to complete the exploration and construction of temporal knowledge graph for grid infrastructure material tracking. A grid infrastructure material tracking platform including platform layer, service layer and application layer was built, and the time series knowledge graph was applied in the material supply risk assessment, material supply risk response, material quality supervision and other scenarios. In more than 60 algorithms of Neo4j graph database, this paper mainly calls path finding and search, intermediary centrality, similarity, link prediction related services. In view of the prediction problem of quality defects or hidden dangers of materials, it is transformed into a sequential knowledge graph reasoning problem combining supplier, batch, model and other data.