In order to effectively monitor some faults of in-wheel motor used in electric vehicle and ensure its operation safety， a fuzzy diagnosis method based on rough set （RS） theory and artificial hydrocarbon networks （AHNs） is proposed for detecting faults of in-wheel motor. Aiming at the operation safety of in-wheel motor， vibration signal was analyzed under the influence of the rotating speed and load torque to refine some symptom parameters （SPs） for characterizing the operation state of in-wheel motor. RS theory is used for proposing a discretization method of SPs with the fuzzy processing of the input layer. AHNs is employed to build a classifier， and its intermolecular energy is considered to optimize AHNs algorithm. Then， fuzzy theory is used for establishing the membership functions of the AHNs multi-outputs， so as to build a fuzzy diagnosis model for realizing the fault diagnosis of the in-wheel motor under complex working conditions. Finally， the effectiveness of the proposed method is verified by in-wheel motor bench test.