The surface roughness of milling titanium alloy TC11 was studied. Three prediction models of surface roughness was established, and the fitting of the three models with measurement of surface roughness value were analyzed. Moreover the signal noise ratio S/N and ANVOA were analyzed respectively, and the size of impact of the each cutting parameters on surface roughness and the optimal combination of cutting parameters were gotten. A method of using the finite element to simulate the size of displacement of milling workpiece was proposed, and the contour arithmetic average deviation was acted as the method for evaluating parameters of surface roughness. Simulation results agree with experimental results, demonstrating that the method is feasible and effective.