When a surface experiencing robotic processing to improve its optical performance (such as removing mid-spatial frequencies, localized grinding errors, and regional surface scratches), spindle speed, tool travel speed, pressure, slurry density as well as groove patterns are main factors to influence surface finishes. Based on the desired material removal rate, the Preston equation can provide optimized pressures and velocities between the tool and processed surface. Various groove patterns, however, can hardly predict by the equation because different patterns can cause unique tool deformation and pressure distribution, leading to determine unique smoothing result. In this paper, four typical groove patterns are studied: non-groove, grid grove, annular groove and radial groove with three typical tool types are evaluated by Finite Element Method (FEM) and statistics. Characteristics of these tools and groove patterns are presented in the end of this paper.