Abstract The wavelet transform approach is proposed in this paper to classify various power system disturbances. On encountering various disturbances, the proposed method is useful to categorize these shortfalls into different groups so that operators can decide on strategies to suppress or eliminate them effectively. Different from those Fourier-based transforms, the wavelet transform approach is more efficient in tracking signal dynamics as time varies. The method has been tested through the classification of various simulated disturbances, including voltage sag, voltage swell, momentary interruption, oscillatory transients and flat-tops. Testing results showed the feasibility and practicality of the method for the applications.