A novel method for high-impedance fault (HIF) detection in distribution systems is presented. Using this method HIFs can be discriminated from isolator leakage current (ILC) and transients such as capacitor switching, load switching (high/low voltage), ground fault, inrush current and no-load line switching. Wavelet transform and principal component analysis are used for feature extraction/selection. A fuzzy inference system is implemented for fault classification and a genetic algorithm is applied for input membership functions adjustment. HIF and ILC data was acquired from experimental tests and the data for other transients was obtained by simulation of a real 20 kV distribution feeder using EMTP. Results show that the proposed procedure is efficient in identifying HIFs from other events.