An artificial neural network (ANN) model was developed to predict boundary shear force distributions in smooth rectangular channels and ducts. Currently the designers often obtain these values with the help of semi-empirical methods. In this paper, as an alternative to these methods, a neural network model is presented. The model was trained and tested using 94 experimental data obtained from the works of best known researchers in this field. The proposed ANN model was found to be superior to existing methods for most of the data sets studied.