This paper presents a method to detect and classify the high impedance fault that occur in the medium voltage distribution network using discrete wavelet transform (DWT) and adaptive neuro-fuzzy inference system (ANFIS). The network is designed using Matlab software and various faults such as high impedance, symmetrical and unsymmetrical fault have been applied to study the effectiveness of the proposed ANFIS classifier method. This is achieved by training the ANFIS classifier using the features (standard deviation values) extracted from the three phase fault current signal by DWT technique for various cases of fault with different values of fault resistance in the system. The success and discrimination rate obtained for identifying and classifying the high impedance fault from the proffered method is 100% whereas the values are 66.7% and 85% respectively for conventional fuzzy based approach. The results indicate that the proposed method is more efficient to identify and discriminate the high impedance fault accurately from other power system faults in the system. [1,2]. To mitigate such crisis, some of the conventional schemes such as minimum reactance approach, voltage and current pattern of the systems and the transients associated with these waves were used to locate the faults in the system [3][4][5][6].Researchers have resented a large number of high impedance fault identification algorithms using a combination of computational intelligence methods such as ANN, Fuzzy, harmonic component analysis using Extreme Learning Machine (ELM), decision tree approach, Bayes classification, nearest neighbor rule approach etc. with the signal processing techniques such as Haar transform, Stockwell transform, Fourier and Wavelet transform analysis for identification of fault in the system [7][8][9][10][11][12][13][14][15][16][17]. On the other side, the advancement in information and communication technologies were used by the researchers to transfer the fault data or information using wireless sensor, communication links and phasor measurement units to locate the fault in the system thereby reducing the blackout of the system but these approaches failed to identify HIF in the system [18][19][20]. Among all the methods presented in the literature, the wavelet analysis outperforms for identifying the fault in the system due to its variable window sizing, strong local analysis stability of frequency band, prone to aliasing effect and energy leaking [21]. For these reasons, the wavelet analysis is used in this paper for feature extraction to classify the faults using ANFIS system. The ANFIS was proposed by Jang based on Takagi-Seguno, the performance was superior than the Fuzzy and ANN methods, as it integrates the merits of both by incorporating the neural network approach to fuzzy at each step thereby giving the high prediction accuracy. Due to these merits, this approach has been used as a classifier in various applications such as image processing, biomedicine application and fault in gear system with high accuracy and ...