This study proposes a color feature extraction method combining computer vision and the K-means clustering algorithm to quantitatively analyze the impact of lotus pattern colors on consumers’ purchase intentions. Based on the HSV color space, the primary color features of different lotus patterns were extracted using the K-means clustering algorithm. The SOR (Stimulus-Organism-Response) model was applied to analyze the relationship between these color features and consumers’ emotions (arousal, pleasure) and purchase intentions. The experiment was conducted using the E-prime visual stimulation environment, involving 40 participants in a psychological experiment. The data collected were analyzed using SPSS and AMOS software for reliability and structural equation model validation. The results indicate a significant correlation between the saturation and brightness of the colors in the patterns and consumers’ emotional responses and purchase intentions. Lotus patterns with colors closer to natural states (high brightness, low saturation) were more likely to stimulate consumer interest. This study provides technical support for optimizing the color design of traditional patterns based on computer vision and offers valuable insights into enhancing consumer perception in product design.