There are many coupling parameters in a nonlinear active disturbances rejection controller(ADRC), but the optimal parameters are difficult to be obtained by the method of conventional empirical turning, which affects the control accuracy of the controller. A single-mechanism optimization algorithm is used to set the ADRC parameters that may be the local optimal solution, which can not effectively improve the control accuracy of the ADRC. For this problem, a parameter optimization design method based on the improved shark optimization algorithm is proposed. In order to solve the problem that the basic optimization algorithm is easy to fall into local optimum and converges slow, a hybrid cross mutation strategy and a double population co-evolution mechanism are proposed, which take the ITAE index as the optimization target of ADRC parameters selection and simulation with a two degrees of freedom manipulator as an example. The result shows that the optimized ADRC has less overshoot and higher control accuracy. After adding external interference, the controller can quickly suppress interference, so it has good anti-interference ability which can be used to optimize the parameters of the ADRC in complex nonlinear systems.