Abstract: Due to limitation of the fuzzy clustering algorthm, a new fuzzy clustering algorithm is proposed by incorporating local distance and membership for image segmentation so as to enhance the segmentation accuracy and robustness. Firstly, combining the neighborhood gray information of pixels, the noise image is adaptively filtered to construct a new filtering image, Secondly, the neighborhood distances between local gray values and clustering centers are smoothed for improving the anti-noise capability, and then a weighted membership function, incorporating local and global spatial membership, is emerged to control the tradeoff between boundary and noise, and as a result the new clustering centers is formed. Finally, the obtained segmentation algorithm is carried out on synthetic image, and real images in different levels of noise. The segmentation results not only can reduce the number of spurious blobs and show better visual effects qualitatively, but also demonstrate the higher segmentation accuracy and ARI quantitatively compared with others common fuzzy clustering algorithms.