<p>To address issues such as periodic and repetitive profile errors, as well as the difficulty of modeling nonlinear systems in CNC cam grinding machines, an innovative contour error optimization method based on model-free adaptive iterative learning and repetitive control is proposed in this paper. Initially, a model-free adaptive iterative learning controller is designed, wherein the complex nonlinear single-axis servo system is dynamically linearized using pseudo-partial derivatives, improving system control accuracy. To resolve asynchronous servo tracking errors between the two axes, a cross-coupled iterative learning controller is employed. Profile errors are effectively compensated through feedback, thereby resolving synchronization issues in the dual-axis system. Additionally, a repetitive control mechanism is utilized, leveraging historical machining data to adjust the reference input, further reducing profile errors. The experimental results of cam grinding demonstrate that this method significantly reduces the contour error of the CNC cam grinding machine, improves machining accuracy, and provides a novel solution for high-precision CNC machining.</p>