As a kind of binder phase, the morphological characteristics of calcium ferrite in sinter directly determine the quality of sinter. In order to improve the comprehensive quality of sinter, a quantitative method for the chemical composition of sinter feed and the content of calcium ferrite in different forms is proposed. In this method, RSM is used to design the test scheme at first, and the sinter samples are prepared with chemically pure reagents as raw materials. Then, the calcium ferrite form in the sinter is classified by polarizing microscopy, and the calcium ferrite content of different forms is counted by surface measurement. Finally, a response surface regression model of the chemical composition of sinter feed and calcium ferrite characteristics is established, and the NSGA-III algorithm is used to optimize the response surface model with multiple objectives. The results show that the R2 coefficient of determination of the response surface model is greater than 0.9, and the P value is less than 0.01, indicating that the model has a high degree of fitting and can pass the significance test, which is statistically significant. The interaction between w(Al₂O₃?) and R²? had a significant effect on the calcium ferrite content. The interactions between w(MgO) and w(Al₂O₃) and w(MgO) and R² had significant effects on the contents of calcium columnar ferrite and other calcium ferrite. At the same time, NSGA-III is used to optimize the response surface model with multiple objectives, and it is determined that the raw material ratio of the most ideal form of calcium ferrite in the sinter is w(MgO)=1.8 , w(Al₂O₃?)=1.7 , and R² =2.3 . The predicted values of calcium ferrite in the sintered samples under this ratio are close to the actual values, which verify the reliability of the model.