 The study proposes a novel model named FWE net for extracting flood inundation extent from Sentinel, one SAR images and compares it with three classical semantic segmentation models and two traditional water body extraction methods. The results show that the FWE net model achieves the highest accuracy among all methods, making it a valuable tool for flood emergency response and disaster relief. This article was authored by Jingming Wang, Shushan Wang, FU Tao Wang, and others.