Distribution systems are constantly at risk of failure due to a variety of factors, including lightning strikes, equipment aging, human error, and breakdown of power system components. These phenomena affect the system reliability and results in expensive repairs, loss of productivity and power loss to customers. Given that faults are unexpected, rapid fault location and isolation are required to minimize their effects on distribution systems. In these research Adaptive network based fuzzy interference system (ANFIS) controller is used to identify the type of faults, location of faults that occurred on the distribution network. The Adaptive network based fuzzy interference system (ANFIS) were trained using various sets of field data. The field data are obtained from the simulation of faults at various points of a transmission line using a computer program based on Matlab. The inputs to Adaptive network based fuzzy interference system(ANFIS) are faulted phase current, voltage and zero sequence of voltage and current measurement available at faulty bus based on Root-Mean-Square values. Neuro fuzzy networks come to be a powerful strategy to develop fuzzy system, since they are capable of learning and providing IF-THEN fuzzy rules in linguistic or explicit form. So, with choosing suitable data for training the Adaptive network based fuzzy interference system (ANFIS) the outputs of Adaptive network based fuzzy interference system(ANFIS) are in form of binary numbers, 1 or 0 for detection of faults and type of fault. The functionality of these system is applied on R1_G5 feeder which is found in Bahir Dar Substation. And different types of fault were studied with respect to R1_G5 distribution network in Bahir Dar. These includes line to ground fault, line to line fault, line to line to ground fault, three phase fault. The basic aim of this research work is to accurately determine type of fault and the exact location of the fault, which is used to provide faster system fault identification, increased reliability and prevent electrical equipment from damaging. As a result, our system generally operates quite well. Subtractive cluster method of generating fuzzy interference system (FIS) with hybrid training of fuzzy interference system (FIS) to locate faults having three inputs of faulty voltage which have Gaussian bell membership function. This ANFIS uses 34 nodes, 18 nonlinear parameters, and 32 linear parameters. 50 parameters and 8 rules are utilized to train 360 dataset. Hence, minimal training RMSE is 1.021625.The simulation results of this research depict that this method is accurate in identifying the fault location and classifying the fault type. Keywords: Adaptive Network Fuzzy Interference System; Fault; Power system protection