Cognitive psychology can help better understand human thinking and behavior, and its application in design can increase the infectious force of design. In this paper, because the brain will generate ERP signals when cognitive stimuli stimulate it, the group sparse decomposition method obtains ERP from EEG signals, and uses the BAP of brain regions to construct the interaction activity pattern of brain regions, and then combines with the interaction genetic algorithm to search for the optimal solution for the data in graphic design by using hierarchical degradation method for processing. The EEG data and interaction behavior characteristics of users were explored through chromatic stimulus response experiments and interaction behavior analysis. It was found that the accuracy of users’ responses to orange, green and purple were all higher than 60%, and P400 in EEG responded to the stimuli most significantly, with P-values of 0.0291, 0.0462, and 0.0355 at P5, P6, and Pz, respectively, with significant effects. The average number of interactions of 46 users in the graphic design enjoyment was 31.47, with the highest success rate of interaction reaching 99.81%, and the average duration of interaction experience for all subjects was 10.26 minutes. Graphic design based on cognitive psychology can capture people’s inner emotions and make better design solutions.