Abstract: In order to realize the intelligent analysis and treatment of mine water inrush and the sharing and reusing of water inrush knowledge in the field, a knowledge base of mine water inrush warning based on ontology is proposed. By summarizing the mechanism of water inrush, the important concepts and their relations in the field of water inrush are summarized. The method of ontology modeling tool Protégé4.3 was used to construct the water inrush body by seven steps to complete the unified and formalized description of the water inrush domain knowledge. Meanwhile, the reasoning rule base is designed according to the expert’s experience and the relevant literature, the ontology model is read by Jena API toolkit and the rule base is loaded to reason out the latent tacit knowledge. Finally, the validity and practicability of this early warning knowledge base are verified by examples of occurrence of floor water inrush. The results show that the prediction result of water inrush from this knowledge base accords with the actual judgment result, which can realize the intelligent discrimination of water inrush danger level.