To deal with the difference of heating strategy resulted from the different conditions of the slabs, a kind of entropy weight-TOPSIS method is proposed, which takes a single furnace zone as the research object and combines with the intermediate process of entropy weight and TOPSIS. A concept of special steel grade level is introduced to quantify steel grade, which is taken as the evaluation indexes with the slab temperature difference, slab location and thickness. This model combines the requirements for the difference of evaluation index in the entropy weight method and the weighted method of TOPSIS. The objective entropy weight is used as the closeness degree of the TOPSIS method, and by normalizing it the final weight of the slab is obtained and the temperature of control zone can be set. The simulation results show that, compared with the fixed weight model, the average and maximum temperature difference are decreased by 5.05 ℃ and 7.77 ℃, and the fluctuation value of furnace temperature is reduced by 8.98 ℃.