Safety management plays a vital role in blasting operations, and blasting safety is closely related to the processes of drilling, blasting, loading, transportation, and dumping, with significant interactions among these procedures. However, due to the diverse sources and complex structure of current blasting safety data, the lack of systematic integration poses challenges for on-site personnel to accurately acquire critical safety knowledge under complex working conditions. To address this issue, this study applies a BERT-BiLSTM-CRF-based method for entity recognition in the field of blasting safety management. The BERT pre-trained model is first used to obtain dynamic word embeddings, followed by optimal label sequence tagging using the BiLSTM-CRF model. A knowledge graph covering seven entity types and nine relationship types is constructed and stored using the open-source Neo4j graph database system. Experimental results show that the F₁-score for all entity types exceeds 60%, demonstrating that the proposed model significantly improves entity recognition accuracy compared to traditional models. Based on this, a knowledge graph-based Q&A system for blasting process safety management in open-pit coal mines is developed, enabling rapid querying of domain knowledge and efficient matching of various blasting processes with safety standards. With the support of this Q&A system, on-site engineers can make timely and informed decisions in complex blasting safety management scenarios.