As a kind of efficient and rapid developed surface processing technique, polishing technology is widely used to improve the surface quality in the precision manufacturing of aero-engine blisk blades. Since blisk blade polishing is a metal removal process which aims to produce smooth surfaces with very low surface roughness, it is difficult to achieve the optimum balance between high material removal rate and low surface roughness. To solve this problem, this study investigates the process parameter optimization to improve the material removal rate with high surface quality in the belt flapwheel polishing of TC4 aero-engine blisk blades. The experiment was designed by Box-Behnken design (BBD) theory of response surface methodology. Then the prediction model of material removal rate was built based on the linear regression analysis of experimental data. To improve material removal rate on the condition that surface roughness meets technical requirements, the single-objective constrained optimization problem was solved utilizing the Particle Swarm Optimization (PSO) technique. Finally, the feasibility of this method was experimentally verified. This study provides a theoretical and experimental reference for the evaluation and improvement of material removal rate with good surface quality in TC4 blisk blade polishing process.