The characteristics of power quality mixed disturbances are analyzed and discussed through the examples. A detailed review of power quality mixed disturbances identification is presented. The research methods of power quality mixed disturbances identification are respectively divided into two major aspects of the single-label and multi-label classifications, and the corresponding current methods and ideas are discussed. For the single-label classification, the five classification ideas, including direct classification, S transform + classifier, wavelet transform + classifier, wavelet transform + neural network, SVM and other methods, are respectively discussed. For the multi-label classifications, the classification idea, strategy, and label correlation are discussed. Finally, the analysis of the current problems of power quality mixed disturbances identification and its prospect are presented.