In this paper, the flow characteristics of the polishing fluid between the polishing pad and the workpiece are studied for the robotic wet polishing process, and the distribution of the polishing fluid radial velocity Ur and the liquid film thickness z at different rotating radii r are revealed. The two-dimensional computational domain consisting of the polishing pad surface, the workpiece wall and the polishing fluid is established. The particle-liquid two-phase flow simulation is carried out in Fluent, and the influence of different rotation rate ω of the polishing pad and different robot swing speeds v2 on the change and distribution of polishing fluid flow rate and temperature are elaborated. The position distribution of the abrasive particles in the wet polishing process and the velocity distribution of particles in the x and y directions impacting on the workpiece surface are simulated and analyzed for polishing fluids with different average abrasive diameters dp. The three-dimensional calculation domain for wet polishing is established; the workpiece surface erosion is simulated in Fluent; the material removal rate MRR and standard deviation of material removal σ on the workpiece surface are calculated considering different combinations of polishing fluid properties Ci and polishing kinematics Pi. Under the same process parameters, the material removal rate test value MRRT and the standard deviation of material removal test value σT are compared with the simulated values, respectively. The results show that under the combination of 64 groups of physical parameters C1-C64 of the polishing fluid, the error between the test value MRRT, σT and the simulationvalue MRR, σ is within 5%. With 64 sets of polishing kinematics parameters P1-P64, the average error between the test value MRRT and the simulated value MRR is 4.19%. However, when the polishing pad rotation rate ω is high, there is an inefficient polishing area in the smaller radius from the polishing pad rotation center, which results in a lower MRRT in some tests than that in simulation, with an maximum error of 8.1%. The average error between the test value σT and the simulation value σ is 3.77%. When the pressure P of the polishing pad is high, the large particles embedded in the polishing pad surface follow its rotation, causing deep scratches on the workpiece surface, which results in a larger σT in some tests, with an maximum error of 7.8%. In conclusion, the material removal principle and the influence of different process parameters in the robotic wet polishing process are revealed in this paper.through modeling and simulation of the particle-liquid two-phase flow, giving an accurate estimation of the material removal rate of the robotic wet polishing process.