Aiming at the problems of periodic and repetitive profile errors and difficult modeling in CNC cam grinding machine,a data driven profile error compensation strategy was proposed.In the single axis servo tracking system of CNC cam grinding machine,a model-free adaptive iterative learning control was added,and the pseudo partial derivative was introduced along the iterative axis to dynamically linearize the complex nonlinear system.In view of the different hysteresis caused by the different servo tracking errors between the two axes,cross coupled iterative learning control was used to feedback the compensation amount to the single axis servo control system according to the cross coupling coefficient to realize the compensation of the cam grinding profile error.According to the simulation experiment results,the proposed profile error compensation strategy can effectively reduce the profile error of the cam and improve the machining accuracy of the CNC cam grinding machine.