Considering the existence of many copper deposits in Iran and the importance of their exploration, mineral potential mapping with high accuracy is an important tool. The process of mineral potential mapping is a cumbersome process which can be performed using different methods. The Hybrid Neuro-Fuzzy approach is one of the powerful ones for mineral potential mapping in which both conceptual and empirical components of earth science information are employed, so using both components simultaneously increase the confidence interval. In this paper we were used Adaptive Neuro-Fuzzy Inference System (ANFIS) for mineral potential mapping in Chahar-Gonbad area 1:100000 sheet, Kerman province. The database consists of geology, geochemistry, airborne radiometric, regional faults, ETM⁺ data, and 22 deposit and occurrence locations. At first, the factor maps were provided in GIS environment in which each cell in the grid data represents a 100 m square on the ground, and then the outputs of this layer were used for training the network. As this technique requires some data for training the network, the occurrence locations were used for training and checking points. Since, the training points were not enough for this procedure, we assigned buffer from 100 to 1000 m for occurrence locations. The results showed that when the buffer is 500 m, the best classification which ANFIS identify about 80% of the known deposits and occurrence locations in high favorability zones.