Abstract: In the smelting process of electric arc furnace (EAF), as the endpoint parameters , the steel temperature, carbon content and phosphorus content affect the EAF's operation and quality of production, and therefore the endpoint control is of gr eat importance. Based on the idea of increasing model, a case based reasoning (CBR) based endpoint prediction model is proposed. In order to minimize the severe nonlinear correlation among the input parameters, and to improve the accuracy and robustness of the model, the result of CBR is corrected by fuzzy least square support vector machines (FLS-SVM). The accuracy of prediction model is remarkably improved and the simulation results demonstrate the efficiency of the method.