The reduction index of sinter is an important parameter affecting the production of blast furnace, and the sintering raw materials and their ratio are one of the fundamental reasons for determining the reduction performance of sinter. In this paper, the response surface method is used to optimize the reduction index of sinter, and the CCD module is used to design the ore blending scheme of sinter. With w(Fe₂O₃)/w(Fe₃O₄), binary basicity [w(CaO)/w(SiO₂)] and w(MgO)/w(Al₂O₃) as variable factors, the reduction index of sinter is taken as the target index. The response surface model between target index and variable factor is established. By fitting the response surface model through experiments, the prediction equation of sinter reduction index is obtained. According to the optimal ratio of raw material composition determined by the prediction model, the prediction model is verified. The results show that the significant effects of the three raw materials on the reduction index of sinter are as follows: binary basicity [w(CaO)/w(SiO₂)], w(MgO)/w(Al₂O₃), w(Fe₂O₃)/w(Fe₃O₄), and the interaction of any two factors among the three is obvious. When the reduction index of sinter is optimal, the raw material ratio is: w(Fe₂O₃)/w(Fe3O4) = 0.76, binary basicity [w(CaO)/w(SiO₂)] = 2.23, w(MgO)/w(Al₂O₃) = 0.82. Through the sintering test, the predicted reduction index value of sinter is 92.41%, the test value is 92.1%, and the content of acicular calcium ferrite in sinter is more, and the microstructure is better, reflecting that the optimal ratio scheme determined by the response surface method is accurate and feasible.