Abstract: Color images of product convey the consumers’ color demands through emotion cognition.In order to investigate the correlation between the color images and color demands,in this paper,a color decision system based on hybrid intelligence method was developed,which adopts the multi-users’ color images as the decision factor and aims at grasping the emotion image of the consumers’ color demands.An improved particle swarm optimization algorithm was proposed using self-adaptive inertia weight factor and chaotic learning factor.The prediction model was built using the improved algorithm to optimize the radial basis function artificial neural network.The consumers’ color demands were obtained by the prediction model.The determined colors for product designing were extracted from the consumers’ color demands using fuzzy C-means cluster method.The test and evaluation results of the automobile exterior color show that the proposed color decision system has an excellent performance with high user satisfaction,and can provide a theoretical guidance for color design of all kinds of products.