Abstract: In order to develop an adaptive model for automatically generating rolling schedules that comply with practical operations in heavy plate rolling, the techniques of Case-Based Reasoning (CBR) and Knowledge Discovery in Database (KDD) are introduced in the modeling process. In the model CBR not only is used for storing and retrieving the cases, i.e. schedules by which high-quality products have been rolled, but also generates an initial schedule for the material to be rolled according to the similarity in attributes between the material and the stored cases. KDD is introduced to find the rules from the operational data records to modify the initial schedule if there is difference between the ongoing rolling practice and the case applied. The experimental results of comparing the schedules against the practical rolling data show that the schedule generated by the new model is more reasonable and conformable to the actual rolling practice than that generated by traditional methods, which indicates that the method proposed in this paper is a promising method for rolling schedule generation modeling.