In view of the complex identification and inaccurate disturbance location of distribution network transient power quality, a disturbance self recognition and localization method based on S-transform matrixes and classification rule tree is proposed. The method analyzes signals by the S-transform and extracted four characteristic quantities from them, then the classification of disturbance signals is performed by use of a rule-based decision tree; and then the disturbances' amplitude change and start-stop time are detected using the amplitude envelope algorithm and amplitude square sum algorithm of S-transform matrixes. The simulation results show that the method is fast and accurate to identify and locate the disturbance signal with low complexity.