Flow over a side weir is one of the more complex flows to simulate in one-dimensional unsteady flow analysis. Various experiments have been applied, but no agreement is apparent in the literature about the best method. In this study, an Artificial Neural Network model has been used to extract a discharge equation for side weirs which accurately estimates overflow discharges. The proposed methodology gives the advantage of accounting for both the geometric and hydraulic characteristics of the overflow structure. The developed model is calibrated and validated using experimental data. Model calibration is achieved by using a Multi-Layer Perceptron (MLP), trained with the back-propagation algorithm. In order to highlight the advantage of the developed model over an existing model widely in use, the model's performance is evaluated according to three comparison criteria. The provided results clearly reflect the ability of the developed model to overcome the weakness of conventional models.