Groundwater has always been considered as one of the important and basic resources of drinking, agricultural and environmental water supply, especially in dry areas. Simulating the groundwater level of a region plays an important role in water resources management. For this reason, today the simulation of the groundwater level using mathematical and computer models with relatively low time and cost is of interest in groundwater studies. In the present study, the groundwater level of Baft area located in Kerman province was simulated using the radial basis function neural network (RBFNN) model. The parameters of precipitation, evaporation, river flow, water demand of the region, amount of abstraction from the aquifer and the level of groundwater with a time delay period as input and the level of the water table in the desired period as the output of the model in a monthly time scale during the statistical period (2002-2016) was selected. Also, in order to evaluate the performance of the model, the statistical indices of egression coefficient (R2), root mean squared error (RMSE), mean absolute error (MAE), mean square error (MSE), normalized mean square error (NMSE) and Willmott’s index of agreement (d) were used. The results of the statistical indicators showed that the radial basis function neural network with R2, RMSE, MAE, MSE, NMSE and d, 0.9989, 0.1256, 0.064, 0.0158, 0.0011, and 0.9997 have a high ability to simulate the groundwater level and provide reliable results.