Abstract: Strip rolling has reached a high level of online detection and automation control capabilities. However, as the proportion of high-end products continues to rise and the demand for the coordinated optimization of multiple indicators such as quality, cost, and energy consumption increases, traditional mechanistic modeling approaches and single-process, single-task rolling process control strategies are insufficient to meet the evolving technological needs of strip rolling. Digital modeling methods and intelligent control technologies, focusing on multi-field, multi-scale, multi-process, and multi-task applications, have rapidly advanced in addressing core quality indicators, including dimensions, performance, and surface quality, as well as production and operational metrics such as efficiency, cost, and energy consumption. The acquisition, feature extraction, and fusion of multi-source heterogeneous data in the rolling process are critical prerequisites for enabling intelligent control. Process determination and anomaly diagnosis have evolved from traditional threshold-based approaches to more comprehensive decision-making and root cause inference, thus raising the bar for the interpretability, traceability, and reusability of diagnostic conclusions. Optimization control, by contrast, needs to achieve multi-objective optimality through coordinated regulation of multiple actuators under equipment capability and process window constraints. Accordingly, with "data acquisition and fusion-quality prediction modeling-quality anomaly diagnosis-multi-objective collaborative optimization" as the overarching thread, this paper systematically reviews recent advances and industrial practices in strip rolling, covering digital modeling, process monitoring and diagnostic decision-making, and optimization control, thereby providing a rigorous reference for the development of intelligent rolling technologies.