Rockburst is one of the most important problem during the excavation of deep buried tunnels. How to predict rockburst accurately is one of the major subjects in geotechnical engineering. With the randomness and implicit vagueness of rockburst in deep buried tunnels, rockburst can be considered as a risk. Fuzzy probability risk theory based on reliability is introduced to evaluate rockburst risk, a new rockburst risk prediction model of fuzzy probability risk analysis is established. This method overcomes the disadvantages of traditional methods, can predict both the rockburst and its probablity. On the basis of Russenes method, the probability density function and membership function are deduced. The proposed method was applied to the rockburst risk of a typical tunnel in western route of the South North Water Transfer Project, the obtained results are compared with traditional methods. Because of its considering of the uncertainties, it provide a new way for the prediction of rockburst.