The invention discloses a system health state prediction method based on Bayes network and base data, comprising steps of systematically analyzing the construction of the relation of nodes of the Bayes network, establishing a prediction model of basic level nodes, and determining a transmission relation of information between nodes, and predicting the health state of the system. The system health state prediction method based on the Bayes network and base data establishes the qualitative and quantitative connection between the basic level data and the system state, solves the problem of system state prediction under the condition that the system level data is hard to determine or hard to obtain, and can fuse the information of various basic levels of the system in order to obtain the prediction result of the system state.