Abstract: To overcome the instability of the diagnostic reasoning results caused by the difficulty in knowledge acquisi- tion,the single knowledge representation,and the poor self-adaptation ability of fault diagnosis reasoning method in tra- ditional hoist fault diagnosis systems,the hoist fault diagnosis method based on knowledge engineering is investigated. Fault diagnostic rule knowledge acquisition methods based on improved attribute importance is proposed,and it pro- vides a data basis for hoist fault diagnosis. The mine hoist fault diagnostic ontology knowledge base is constructed and the fault diagnostic ontology knowledge representation methods based on OWL DL and fault diagnostic rule knowledge representation methods based on SWRL are proposed,and the hoist system structure and the diagnosis knowledge inte- gration are implemented. The probability of the ontology knowledge is extended,and a new fault diagnosis uncertainty knowledge reasoning method is proposed,which are based on ontology and Bayesian. Based on the theory and method above,the fault monitoring and diagnosis system of the mine hoist is developed,and the method is proved to be feasible and reliabile.