Abstract: In this paper, wheat bran was used as raw material.Based on the previous and single factor experiments, the amount of wheat bran powder, the dosage of gel, the amount of CMC were used as the independent variable and the sensory comprehensive score were selected as the response value.According to the response Box-Benhnken test, the prediction principle of quadratic polynomial regression equation was simulated by the analysis of three factors and three levels, and the preparation process of high fiber filling material was optimized. The results showed that the model was of great difference, and the established regression equation for sensory score of high dietary fiber had an excellent goodness of fit.Therefore high dietary fiber could be analyzed and predicted by the model. Factors influencing sensory score of high dietary fiber were in the order as follows: gel dosage > amount of wheat bran powder > CMC dosage.The response surface plots showed that the interaction between amount of wheat bran powder and dosage of gel was extremely significant and those between amount of wheat bran powder and CMC dosage was also extremely significant, while the interaction between dosage of gel and CMC dosage was significant.The optimum technological conditions were wheat bran powder dosage of 38 g ( wheat bran∶ starch = 3∶ 1, 120~140 mesh of size) , gel dosage of 0.75 g, CMC dosage of 1.4 g. Under these conditions, the product was evenly coated, the coating thickness was suitable, the shear stress was 1185 g/cm2, the water holding capacity was up to 98.52%, and the sensory score was as high as 94.67, which was highly consistent with the predictive value.The conclusion was drawn that the results of this study would provide a reference for the research and development of high fiber fillings.