Signal processing is a method used in electrical engineering to analyse and modify signals hence improving the quality. In power quality classification, signal processing is employed as an alternative for better detection of disturbances. This is done with the application of some extraction features to amplify the differences of each disturbance. However, due to border distortion effect, signal processing results usually encounter problem. Border distortion is the aberration in results due to tempering, appearing at the starting and the ending of the waveform when signals are processed and modified using signal processing techniques. This leads to misclassification of PQ disturbances, causing problems in determining the correct mitigation technique to be taken to reduce power quality disturbances. In this paper, minimization of border distortion of features extraction resulting from Stockwell transform will be discussed via the implementation of Hanning Taper and polynomial fitting. The discussion involves the three types of multiple power quality disturbances which are voltage sag with harmonics, voltage sag with transient, and harmonics with transient. The modification technique in Stockwell transform will be analysed to prove the accuracy and the consistency in features extraction. In this study, the border distortion effect was successfully reduced by an average of 88.704%.