Abstract: In the cold rolling process, owing to the absence of effective process monitoring methods, it is challenging to detect and optimize the degradation of controller performance promptly, often resulting in a reactive approach to anomaly remedi-ation and negatively impacting production efficiency. To address this issue, a novel mechanism-guided maintenance strategy was proposed for thickness and tension control in tandem cold rolling. Specifically, a forward-looking index was introduced, and the mechanism-guided Hurst (MH) index was developed by incorporating domain knowledge, enabling real-time monitoring of the production process. When controller performance falls below expectations, the MH index guides the whale optimization algorithm (WOA) to achieve multi-controller coordinated optimization and overcome the global optimization challenges arising from the nonlinear, multivariable, and strongly coupled nature of cold rolling. Simulation results demonstrate that this approach effectively monitors and enhances the performance of multiple con-trollers in cold rolling. Furthermore, the method was successfully implemented in an industrial setting, enabling real-time monitoring, with enhanced control accuracy post-optimization, providing valuable insights for industrial applications.