In order to improve the flatness of cold rolled strips, the accurate knowledge of actuator performance in flatness control is required. As the basis of flatness control, efficiencies of flatness actuators provide a quantitative description to the law of flatness control. In this paper, a 1250 single stand 6-H reversible UCM cold mill was taken as the object of this study, with efficiency factors of individual actuators analyzed for better flatness control. For the purpose of obtaining accurate efficiency factors matrixes of actuators, a self-learning determination model of actuator efficiency factors was established in accordance with the practical rolling processes. The actuator efficiency factors can be determined and improved continuously by the self-learning model with correlative measurement flatness data input. With this study, performance of actuators in flatness control of UCM cold rolling mill has been obtained, which provides a theoretical basis for better flatness control.