In this paper, a new method based on rough-fuzzy set and differential immune clone clustering algorithm (DICCA) for image segmentation is proposed. By replacing hard clustering with fuzzy clustering through incorporating rough-fuzzy set into DICCA, this algorithm can obtain more abundant clustering information. Specially, as the advantage of rough set is processing uncertain data, the proposed algorithm is more conducive to solve the uncertainty problem. In experiments, nine images are used for segmentation and four algorithms are chosen for comparison to validate the performance in the clustering stability. The experimental results show that the algorithm has higher segmentation accuracy and better segmentation results.