An enhanced mobile robot navigation method based on Adaptive Neuro-Fuzzy Inference System (ANFIS) in unknown and static environments is presented in this paper. A system's configuration can have many alternatives, such as defuzzification method, number of fuzzy sets, type of membership functions, etc. To figure out a better configuration of the system which mimics human navigation experience better and have better navigation performance, systems using various configurations are trained with MATLAB Neuro-Fuzzy Designer and then tested with considerable environments in simulation. Moreover, an extra strategy for avoiding infinite loops in concave environments is also incorporated in our method. Simulation results show that trapezoidal function and more fuzzy rules lead to better training and navigation performance. The optimal configuration works well in various complicated environments, performs better than other papers and successfully escaped concave obstacles.