In this study, we propose a method for designing an illuminant for supporting color discrimination with maintaining the arbitrary illuminant color. It was applied to the problem of foreign substances detection in blueberry jam because it is especially difficult and essential for quality assurance. To obtain the optimal illuminant, a spectral dataset of metameric illuminants was constructed and a combination of the metameric illuminants with a given chromaticity was derived to maximize the color differences between the jam and foreign substances (leaf, branch, and stone). A color appearance model (CIECAM02) was used to evaluate difficulty of the substance discrimination. The designed illuminant was implemented using a synthesized spectral light source, which clearly showed that the color discrimination of jam and foreign substances was enhanced. The application field for this technique is not limited to foreign substance detection in blueberry jam because it could also be applied to quality control processes of other industrial products, medical imaging, and so on.