Abstract: Because the conventional proportion-integral-derivative (PID) algorithm has the shortcomings of low accuracy and poor adaptability, a composite method, which includes the TS fuzzy neural network (TS-FNN) and PID controller, is proposed. This control strategy can improve the accuracy of dissolved oxygen (DO) concentration by the self-learning ability of TS-FNN. Meanwhile, the parameters of the controller can be adjusted on-line by constructing performance coordination factor. Then, this method is tested based on the international benchmark simulation platform.Resultsshow that the proposed method can achieve better dynamic performance, compared with the conventional back-propagation (BP) controller and PID controller.