Abstract: Fuzzy C means （FCM） clustering can be used for image segmentation,which is a procedure of the label following a unsupervised fuzzy clustering.However,there are some deficiencies in the FCM algorithm,which limit its application in some aspects.A fuzzy clustering based approach to image segmentation is proposed in this paper,which conquers the deficiencies encountered in FCM.The feasibility of the approach is proved mathematically and experimentally.