In this study, an experimental investigation was carried out to analyze the thermal mixing phenomena in a rectangular cross-section narrow channel. The channel has two circular water jet inlets, in different temperatures and a circular exit hole to supply continuity of mass. Several parameters were used in the experiments to control thermal mixing efficiency such as inclination angle of the channel, ratio of flow rate of inlet fluid, temperature difference between hot and cold jets and jet inlets diameters. Thermal mixing index was calculated from the measured temperatures. In order to reduce experimental time, an artificial neural network (ANN) model was built with limited number of measurements for a forward model. The obtained results indicated that estimation data of ANN shows that the ANN model can predict the output parameters without carry out any experiment. (C) 2013 Elsevier B.V. All rights reserved.