Abstract: A diagnostic method on rotating machinery based on fuzzy C-means clustering and rough set theory is proposed in this paper, which includes two processes: one is rough-set-technique-based diagnostic rules requisition and the other is diagnosis of new objects based on the rules obtained. The diagnostic rule requisition process takes into account the duplicated and conflicting objects in decision table and makes the rules obtained covering all the learning objects, and obtains corresponding rules strength. In the diagnostic process, the significance of condition attributes of rales, the matching degree of condition attributes of objects with diagnostic rales, the strength of those rules and the conclusion threshold is considered to give diagnosis conclusion and belief degree of conclusion. So the conclusion based on the proposed method is objective. In the end of this paper, this method is employed to diagnose the faults of rotating machinery and the result is proved to be effective.