By the intelligent strategy of dynamic optimization in flatness target curve, the potential of the flatness control system can be completely fulfilled and the satisfying strip flatness can be speedily obtained. The traditional intervention pattern is static, open‐loop, non‐intelligent. In this paper, a new dynamic adjustment model of flatness target curve has been designed, by which the initial setpoint can be real‐timely regulated. The closed‐loop control system, equipped with the function of intelligent optimization and dynamic regulation, has been developed. A hybrid intelligent algorithm called ZOGE, in which both the searching definiteness and randomness are taken into account and the multipoint searching and single‐point searching are simultaneously carried through, has been proposed. The equivalent factor of flatness target curve has been optimized by function module HYBALO, which is developed on the basis of the hybrid intelligent algorithm and is operated on CFC configuration language in SIMATIC TDC controller. The testing in the field has been conducted through a 1450 mm 5‐stand tandem cold mill. The preliminary finding from the main results of testing in the field is reported: by means of the intelligent compensation model of flatness target curve, the content flatness quality of strip with thin thickness in slow rolling speed is effectively guaranteed. The maximum compensation efficiency of speed is 72.19%. The maximum compensation efficiency of thickness is 98.37%. The ability of eliminating strip flatness defect in dynamic regulation mode of flatness target curve is more powerful than the ability of eliminating strip flatness defect in conventional regulation mode.