Objective Aiming at the stability of the actual trajectory tracking of the flexible joint manipulator with the uncertainty of external disturbances a method combining adaptive dynamic surface control and a neural network was proposed. Methods According to the characteristics of the radial basis function RBF neural network the functions and unknown parameters in the nonlinear system were approximated and the interference item from the outside to the system was compensated by designing the damping item. According to the knowledge of the dynamic surface the controller in the nonlinear system was designed and the joint trajectory tracking control was realized. Results The simulation results show that this method can overcome the disturbance uncertainty item in the nonlinear system and achieve a better tracking effect of the connecting rod rotation angle q of the manipulator and the error is reduced within 5%. This method has strong tracking stability. As time goes on the tracking error becomes smaller and tends to 0 and the estimation and approximation of the parameters have reached ideal thresholds. Conclusion This method ensures the semi-global stability of the closed-loop nonlinear system the parameter adjustment can be utilized to achieve arbitrarily small tracking errors and the designed controller not only ensures the position tracking stability of the manipulator but also solves the problem of tracking jitter well.