To classify power quality disturbances， a method using 2-dimensional discrete static wavelet transform is introduced. First， the technique of 2-dimensional wavelet transform（WT） is employed to extract the energy distribution features of the distorted signal at three different high-frequency and one low-frequency. Then， the probabilistic neural network（PNN） classifies these extracted features to identify the disturbance type according to the energy features and module maximum value at the horizontal high-frequency. It is convenient for using the best simple wavelet function such as db1 and decomposing one-level WT. The simulation results show the effectiveness of the proposed method.