Biochemical Oxygen Demand is one of the most crucial water quality parameters to assess of water pollution of rivers. Nevertheless, BOD needs longer periods (5 days) to get results. The objective of this research is to build a computational model based on the artificial neural networks, including Multilayer Perceptron Network (MLP), and Radial Basis Function network (RBF) for simulating BOD5 in the lower Sai Gon – Dong Nai rivers, and to evaluate the simulation efficiency between MLP and RBF. Seven different input combinations were constructed using Pearson correlation coefficients between each water quality parameter (COD, DO, TSS, Coliform, P–PO4 3–, T, and N–NH4+) and BOD5. Five years (2013 to 2018) of monthly data from eight water quality monitoring stations within the study area were compiled, which were divided into two sub–sets (ratio 75:25) for model training and model testing. The results indicated that both the models satisfactorily simulated BOD5, but the RBF model with the combinations of variables numbered 5 (COD, DO, TSS, Coliform, P–PO4 3–) demonstrated the best performance, values of Nash–Sutcliffe efficiency (NSE), coefficient of determination (R2), and root mean square error (RMSE) were 0,848, 0,865, and 0,454, respectively. The results of this research are also the foundation for short–term prediction of BOD5, as well as the simulation of the other water quality parameters in the area.