Abstract: In view of problems of limited knowledge content, difficult knowledge fusion and sharing, and inability to mining related knowledge from unstructured data existed in coal mine roadway support expert system, a knowledge graph of coal mine roadway support field was constructed. Firstly, knowledge of coal mine roadway support field is modeled by designing field concept, relationship and attribute. Then, knowledge is acquired from structured, semi-structured and unstructured data sources of coal mine roadway support field, and entity recognition is carried out based on deep learning model BI-LSTM-CRF. Finally, knowledge graph of coal mine roadway support field is formed by using graph database Neo4j to store knowledge of coal mine roadway support field. The knowledge graph of coal mine roadway support field can further improve efficiency of roadway support design and management, and provide knowledge support for intelligent management of coal mine roadway support.