A color image segmentation algorithm based on disjunctive normal level set（ DNLS） is proposed to improve the result of color image segmentation. The color image is converted to HSV color space for taking more effective pixel information in the image. An energy model based on disjunctive normal level set is proposed

which is based on the weights of the different channels

and the energy function of the level set is rewritten. The evolutive polytopes are employed to simulate the foreground region in the image

which avoids the difficulty that re-inialization or added regularized term during the evolutive process of the traditonal level set. Compared to other popular algorithms quantitatively in the image segmentation experiments

validity and competitive performances of the algorithm are demonstrated.