Abstract: Tunnel water and mud inrush disasters pose major geological risks to construction safety.Their formation involves complex mechanisms,high suddenness,and strong multi-factor coupling,which limit knowledge accumulation and reuse in experience-based assessment and treatment decisions.A multi-factor coupling ontology-based method for modeling and knowledge graph construction was developed for such disasters.Interactions among disaster-forming environment,causative factors,and affected objects were analyzed using disaster system theory.A three-level ontology comprising event,object,and attribute layers was constructed to achieve semantic definition,hierarchical organization,and relational constraints of disaster elements.Unstructured case texts from the literature were processed using the YEDDA annotation tool and a BERT-BiLSTM-CRF model for high-accuracy entity extraction.Predefined relation schemas and a human-in-the-loop validation mechanism were applied to establish semantic links among entities and to build a structured triple-based knowledge base.A knowledge graph with semantic consistency and causal representation was implemented using the Neo4j graph database and the Cypher query language.The approach enables standardized integration and intelligent utilization of disaster knowledge,as well as multi-dimensional and multi-hop retrieval and reasoning over cases.Results show that the method effectively integrates multi-source heterogeneous knowledge and supports rapid query,assessment,and treatment decisions for tunnel water and mud inrush disasters,providing a reusable framework for engineering-oriented knowledge graph construction.