This article attempts to put the science of color preference on a firm theoretical footing. A model of color contrast, in which the perceived color of an area is influenced by the surrounding colors, is proposed. It is based on double opponent cells, neural units which respond preferentially to one of the four opponent colors, blue, yellow, red, and green. It is shown that by modeling the receptive fields (RFs) of these cells, one can account for a number of color contrast effects. These include the direction of the color contrast shift, the magnitude of the shift as a function of the extent of the surround, and the self-inhibitory nature of a colored area. It is then demonstrated, with an additional assumption relating mean activation of the double opponent units to liking, that the same model can be used to advantage in understanding color preference. It is first shown that the model predicts that preference should increase with increasing saturation. Next, it is shown that the inhibition of a fully saturated field on a less saturated field is greater than vice versa, in keeping with preference results in priming experiments. The model is then shown to predict that complementary colors are in maximum harmony, also in accord with observed results. Finally, the concept of virtual saturation is introduced in order to make predictions regarding hue preference.