Aiming at the problem that traditional fuzzy c-means (FCM)clustering algorithm is sensitive to noise in image segmentation, an extreme filtering fuzzy c-means with neighborhood constraints (EFFCM_N) clustering segmentation method is proposed in this paper. The algorithm first filters the image, and then uses the processed image for fuzzy C-means clustering segmentation. In addition, the objective function is constrained by adding neighborhood information. Through the segmentation experiments of synthetic image and natural color image, the experimental results show that the proposed EFFCM_N algorithm is obviously superior to the traditional FCM algorithm and some improved algorithms in segmentation effect and quality, and has stronger robustness to various noises.