In order to quickly find the cause of the spindle failure, inverse Petri net was adopted to diagnose the fault of CNC(computerized numerical control) machine tool spindle. An inverse Petri net model of related failure for CNC machine tool spindle was established to obtain the inverse Petri net reachability graph of the spindle. And the failure of subsystems, failure modes and failure causes can be thorough analyzed. In addition, the relevance of fault modes and fault causes can be analyzed by the reachability graph. By analyzing state transferring reachable set, we can grasp the transmission process of spindle related failure and get the root cause of failure. By taking the loud noise of spindle as an example, its fault diagnosis process with inverse Petri net was described. Studying inverse Petri net of spindle related failure can help to diagnosis and repair the spindle quickly, shorten the repair time and improve the availability of the spindle.