Aiming at detecting cloud interference over urban area, an algorithm in this research is proposed to detect urban cloud area combining extracting edge information with information entropy, focusing on distinguishing complex surface features accurately to retain intact surface information. Firstly, image edge sharpening is used. Secondly, Canny edge detector and closing operation are applied to extract and strengthen edge features. Thirdly, information entropy extraction is adopted to ensure cloud positional accuracy. Compared with traditional cloud detection methods, this algorithm protects the integrity of urban surface features efficiently, improving the segmentation accuracy. Test results prove the effectiveness of this algorithm.