Color dot painting is an art technique that visually constructs an image from a large number of small pixels, and the number of pixels directly affects the cost of composition. The optimized point selection and composition method provides an important way to achieve low-cost printing. At present, there are difficulties in point drawing generation, such as difficulty in evenly distributing multi-channel sampling points, difficulty in balancing color levels, and significant computational costs. This article proposes a method for generating color dots based on superpixel adaptive clustering and linear programming optimal points. This method uses an adaptive clustering method based on color density peak to obtain the optimal number of clusters on the basis of preliminary superpixel image segmentation, and further divides into sub clusters. Then, the color mean of each sub cluster is used as the optimal interval distance for selecting points within the sub clusters. At the same time, based on the SSIM (structural similarity index measure) index, a target optimization model was established. The mathematical optimizer Gurobi was used to achieve model point selection, which preserves the minimum number of points while maintaining uniform distribution and color gradient levels within the clusters, in order to improve the visualization effect of the generated dot image. The experimental results show that the algorithm proposed in this paper greatly reduces the number of pixels and outperforms the current method in terms of evaluation metrics such as mean SSIM and PSNR (peak signal to noise ratio) of generated dots.