The present research aims to develop an Artificial Neural Network (ANN), able to predict in the short term, future water demands. Because of the high precision with time series data, the NARX (Nonlinear Autoregressive with External Input), a type of ANN, was chosen. Several tests were conducted in order to find the best forecast model, and the results showed that the NARX is an efficient predictor of water demand, requiring a good database for the ANN training phase.