This document summarizes an unsupervised change detection method for satellite images using Markov random field fuzzy c-means (MRFFCM) clustering. The method first generates a difference image from multitemporal satellite images using image fusion techniques. It then applies MRFFCM clustering to the difference image to segment it into changed and unchanged regions. Experimental results on real synthetic aperture radar images show that MRFFCM clustering produces more accurate change detection results with less error than previous approaches, while also having lower time complexity. The method is evaluated on datasets from Bern, Ottawa, and the Yellow River region, demonstrating its effectiveness.