Rock burst in deep level is one of the major problems troubling the mining process. Prediction of rock burst proneness requires comprehensive analysis of multiple factors. In this paper, the proneness of rock burst is considered as a domain of fuzzy set. Using the indirect fuzzy identification method, a method in fuzzy mathematics, the proneness of rock burst is studied through fuzzy classification, as the degree of the rocks fuzzy subset appropinquity to the subset of types of rock burst. It proved to be more objective and scientific compared with purely synthetic fuzzy judgement according to expert's evaluation.