Abstract: Fuzzy C-means (FCM) algorithm only considers the gray information of pixels and ignores the neighborhood information of pixels, resulting in inaccurate segmentation results. To solve this problem, considering the distribution characteristics and interaction between image pixels, this paper designs a complexity to increase the weight of pixel spatial neighborhood information in the algorithm. This complexity information was integrated into FCM algorithm. Combined with intuitionistic fuzzy set theory, hesitation degree and non-membership degree were introduced to improve the uncertain information in the image and optimize the membership matrix. Experimental results show that the algorithm weakens the influence of noise on the image and has stronger robustness to the processing of edge details.