Trajectory Changing of the industrial robots are often accompanied by system noise, interference, and the introduction of its own inertia parameters change, and the traditional iterative control algorithm is difficult to achieve high precision and high-speed control requirements. This article combines adaptive control and robust control and the iterative algorithm together to improve the control precision; when the given task is changed, the historical control experience estimates the changes in training estimated input after the system's expectations, as the initial iteration controller input, the controller works as joint fuzzy cerebellar feed-forward control, when new tasks arise it avoids the choice of the initial amount of the blind in the order to achieve high-speed control purposes. At last the robot system simulation results show the validity and rationality of the algorithm.