In this study, Site Groundwater Rating (SGR) in Amirkabir tunnel has been predicted using Artificial Neural Networks. SGR is the first rating method which by considering the parameters like joint frequency, joint aperture, schistosity, crashed zones, karstification, soil permeability coefficient, tunnel location in water table or piezometric surface, and the amount and intensity of annual raining ... [Show full abstract] of the area, classifies the tunnel path from the risk of groundwater seepage point of view. In this article, using a multilayer perceptron neural network, prediction of SGR along Amirkabir tunnel path was performed. Field data obtained from primary studies in the tunnel was used to train and test the prepared network. For the testing set, modeling results showed that SGR could be predicted with mean error of 3.16% using a perceptron network including a hidden layer. High correlation between SGR of the tunnel path and the network answers, confirmed our prepared Artificial Neural Networks.