Abstarct: Abstract High Impedance Fault (HIF), which can cause fire and damage, is not detected by conventional overcurrent relays because of low current magnitude. It is critical to distinguish it from other disturbances such as switching on/off the capacitor bank, linear and nonlinear loads, distributed generation (DG), induction motor, and inrush current caused by distribution transformer energization. So far, various techniques have been proposed to separate HIF from other disturbances, but most of them don’t consider the effect of DGs. This thesis presents a method baxsed on the fast discrete orthonormal Stockwell transform (FDOST) to detect HIFs from other disturbances. The proposed disturbance detection index is defined as the ratio of Shannon entropy of the real part of the aforementioned transform coefficients of the disturbance current signal energy to its Kurtosis. This method only needs half a cycle of post-disturbance current, and the thresholds used to detect disturbances are independent of the studied network and are always fixed values. The proposed method is tested on IEEE 13-node and IEEE 34-node test feeders, considering DGs and noise. Also, the effectiveness of the proposed method for the high impedance fault recorded on a real feeder in Iran is also evaluated. The simulation results show that the proposed method has more than 98.9% accuracy without DGs and more than 98.4% in the presence of DGs under normal and noisy conditions. Furthermore, this method has a low computational burden and the average detection time is nearly 3.5 ms.