With the increasing demand for automation and efficiency of the power grid, the power industry needs to accelerate the pace of building a highly intelligent and intelligent platform. However, the power operation and maintenance knowledge base for the professional model field is different from the common knowledge bases on the market. Some conventional application technologies cannot be used directly. The technical difficulties and practical business problems faced by the power operation and maintenance knowledge base in real scenarios are still not very good. solution. This paper uses technologies such as natural language processing, combined with deep neural network models, to construct a set of portable and easy-to-retrieve power knowledge graphs through knowledge extraction, knowledge fusion, knowledge reasoning and other technologies.