The glass-forming region in the Sm-Si-Al-O-N system at 1700°C was studied by using an artificial neural network (ANN). An artificial neural network (ANN) was trained and tested with fifty experimentally determined examples in the investigated system with a back-propagation algorithm. The results of all forty-two examples of the training set and the testing set of eight obtained from the ANN are in good agreement with experiments. The glass-forming boundaries on the planes of 25, 27.5, 30, 32.5, 35 and 40 eq% N were predicted with the trained network, which indicates the tendency of contraction towards the Si-rich region with increasing nitrogen content. The approach shows a considerable promise for applications to determination of nonlinear boundaries of glass-forming regions in complex material systems, in cases where thermodynamic calculations of phase equilibria are not effective.