The frequency of roof-caving accidents ranks first among all coal mine accidents. However, the scattered knowledge system in this field and the lack of standardization exacerbate the difficulty of analyzing roof fall accidents. This study proposes an ontology-based semantic modeling method for roof fall accidents to share and reuse roof fall knowledge for intelligent decision-making. The crucial concepts of roof fall accidents and the correlations between concepts are summarized by analyzing the roof fall knowledge, providing a standard framework to represent the prior knowledge in this field. Besides, the ontology modeling tool Protégé is used to construct the ontology. As for ontology-based deep information mining and semantic reasoning, semantic rules based on expert experience and data fusion technology are proposed to evaluate mines’ potential risks comprehensively. In addition, the roof-falling rules are formalized based on the Jena syntax to make the ontology uniformly expressed in the computer. The Jena reasoning engine is utilized to mine potential tacit knowledge and preventive measures or solutions. The proposed method is demonstrated using roof fall cases, which confirms its validity and practicability. Results indicate that this method can realize the storage, management, and sharing of roof fall accident knowledge. Furthermore, it can provide accurate and comprehensive experience knowledge for the roof fall knowledge requester.