In computer numerical control (CNC) machining, ensuring the production of high-quality of machined surfaces is important alongside factors such as high production speed and precision. However, existing surface quality evaluation methods are inconclusive, being largely based on photographic or human visual inspection. Therefore, a method must be developed to quantify and evaluate the surface quality of machined parts, enabling the further study of high-quality machining conditions. Improvements in the surface quality of machined workpieces and identification of parts having low surface quality is expected to reduce the need for and improve the efficiency of hand polishing processes. Therefore, this study evaluated the surface quality of a machined surface by associating an index of luminance measured by a vision system with an index of surface roughness measured by a microscope. Furthermore, the luminance differences were imaged to visualize the surface quality. Different conditions of pick feed and feed amount were used, and the method of imaging them was studied. It was observed that low values of Lq/Lm (root mean square luminance/luminance mean) and Ldq (root mean square luminance difference) were reliable indices of high quality. The surface quality was visually evaluated, and the uneven machined surface was clearly identified by the obtained luminance difference images using a threshold value.