Effluent quality of vertical tube biological reactor (VTBR) was simulated by support vector machine. Average R2 of the training set and the test set were used to balance the fitness and the predictive ability of SVR models, and evaluate the general performance of SVR models. Following a special grid search procedure, optimal parameters were determined for these SVR models. Results indicted that these SVR models have good fitness and predictive ability. The developed model may be integrated into an advanced control system for the VTBR using different control strategies with further work. Furthermore, SVR was found to be a useful and promising tool that is worth consideration in the prediction of effluent of wastewater treatment processes.