Color, as one of the most important visual features in a product system, can easily cause psychological and emotional resonances. Depending on the importance of user's emotional needs in the market and the complexity of a mining process, it can be reflected in the adjective description of color image. In this study, ontology combined with big data was used to mine the color image, a visual sense of user's psychological requirements. The results showed that the constructed color image ontology model could logically express the conceptual knowledge of color image. And this method solved problems that exist in current research on image mining and at the same time increased the accuracy of image information mining. Further examination of the color brand images was used to obtain the idea of "Stylish" and "Natural." Finally, the acquisition of their physical form was named and surveyed to prove the feasibility of the research approach. This is a new attempt to apply ontology to the color image cognition process.