Vehicle shock absorber vibration is classified and analyzed. It shows various Gaussianity and symmetry under normal or abnormal sound. The theory of bispectra analysis is reviewed and vibration signal selected bispectra for vehicle shock absorber are plotted. A new method of fault feature extracting for shock absorber vibration based on the selected bispectra. Thus achieving the diagnosis for the abnormal sound of the vehicle shock absorber by the means of the discriminance of mahalanobis distance. The faults diagnosed accord well with the actual situation, shows that method presented is of practicality and reliability. It is useful for understanding the mechanism of shock absorber abnormal sound and for its control.