To predict the current efficiency of the aluminum reduction process, an intelligent integrated prediction model is presented. Firstly, in view of the deficiency of the fuzzy c-means clustering and unsupervised clustering method, an improved fuzzy c-means supervised clustering algorithm is proposed and on the basis of it, a supervised distributed support vector machine is constructed for the prediction of the current efficiency. Secondly, based on the aluminum electrolytic reaction principle, the mechanism model of current efficiency is established. Finally, by the weighted integration of the two models, the intelligent integrated model of current efficiency is achieved. The simulation results of the field data show that the integrated model has higher precision, and can be used to forecast the aluminum electrolysis current efficiency in practice.