The ??2 gain robust following control method based on fuzzy neural networks for underactuated autonomous
underwater vehicles(AUV) is proposed. The proposed control method can be used to approximate the uncertainty of the dynamic model through the online adaptive learning algorithm. The controller can overcome the influence on following errors of missing sway force. The unknown ocean current disturbance of three degrees of freedom is considered. It is realized that the ??2 gain of fuzzy neural network approximation errors is less than ??. The uniform ultimate boundedness of closed loop system error signals is proved by using Lyapunov stability theory. Finally, simulation experiments with the precise system parameters and parameters perturbation are given, and the results show the good following result and the strong robustness of the proposed control method.