The rapid growth in the field of image processing is compelling for the performanceenhancement in practical applications by processing image correctly. Image segmentationis one of the most important requirements of image processing, which isthe process of partitioning an image into a set of objects and backgrounds. Segmentationplays vital role in analyzing an image automatically. The main objectiveof segmentation is to trace certain objects of interest by ignoring the effect of light,noise and texture on them. Among the different segmentation approaches, graphbased techniques are most popular due to their capabilities of generating goodsegmentation structures. In this paper, we have proposed graph based discreteparticle swarm optimization approach for minimal cut image segmentation. Whichgenerates better segmentation results than other methods as well as helpful in thedevelopment of new hybrid methods.