Flood detection using SAR (Synthetic Aperture Radar) is crucial for disaster response, as floods are among the most severe natural disasters, with increasing frequency and intensity due to climate change and urbanization. SAR offers advantages over multi-spectral sensors by providing all-weather, all-time imaging, enabling flood monitoring regardless of weather or illumination conditions. Recent advancements in SAR data availability have spurred research in flood mapping, with techniques like fuzzy logic, machine learning, and data fusion demonstrating superiority over traditional methods. However, evaluating the performance of flood mapping techniques requires synthetic quality metrics and high-quality reference data. Open SAR datasets with ground-truth information support validation and reproducibility. Despite progress, SAR-based flood monitoring faces challenges in vegetated and urban areas, where complex scattering mechanisms hinder accurate water extraction. This review discusses classification methodologies, SAR datasets, validation strategies, and future perspectives for SAR-based flood mapping. It highlights the importance of backscattering coefficients and interferometric coherence in flood detection, and reviews various processing techniques, including thresholding, fuzzy classifiers, machine learning, and data assimilation. The review also addresses challenging scenarios, such as vegetation cover, urban areas, and shadow/layover regions, where SAR data processing is particularly complex. The paper emphasizes the need for advanced techniques and datasets to improve flood detection accuracy and reliability.Flood detection using SAR (Synthetic Aperture Radar) is crucial for disaster response, as floods are among the most severe natural disasters, with increasing frequency and intensity due to climate change and urbanization. SAR offers advantages over multi-spectral sensors by providing all-weather, all-time imaging, enabling flood monitoring regardless of weather or illumination conditions. Recent advancements in SAR data availability have spurred research in flood mapping, with techniques like fuzzy logic, machine learning, and data fusion demonstrating superiority over traditional methods. However, evaluating the performance of flood mapping techniques requires synthetic quality metrics and high-quality reference data. Open SAR datasets with ground-truth information support validation and reproducibility. Despite progress, SAR-based flood monitoring faces challenges in vegetated and urban areas, where complex scattering mechanisms hinder accurate water extraction. This review discusses classification methodologies, SAR datasets, validation strategies, and future perspectives for SAR-based flood mapping. It highlights the importance of backscattering coefficients and interferometric coherence in flood detection, and reviews various processing techniques, including thresholding, fuzzy classifiers, machine learning, and data assimilation. The review also addresses challenging scenarios, such as vegetation cover, urban areas, and shadow/layover regions, where SAR data processing is particularly complex. The paper emphasizes the need for advanced techniques and datasets to improve flood detection accuracy and reliability.