<p>Flash floods are one of the most deadly and expensive natural disasters in the world, claiming over 5,000 lives annually. In Bangladesh, the combination of topographical features and climatic conditions makes the country highly susceptible to flash flooding, particularly pre-monsoon season in the northeastern area. Therefore, it is essential to conduct A thorough spatial analysis of flash flood risk by comparing MCDM approaches to identify the best technique for delineating flash flood-prone regions and providing effective reduction and adaptation actions. This study employs several MCDM techniques including the Analytical Hierarchy Process (AHP), Fuzzy AHP, and traditional Fuzzy Logic to assess and map spatial flash flood risk in the Sunamganj district of northeastern Bangladesh. Using a total of 14 spatial criteria, flash flood risk maps were generated, and the performance of each MCDM approach was examined through Receiver Operating Characteristic (ROC) curve and Area Under the Curve (AUC) analysis. The results demonstrate that AHP outperforms both Fuzzy AHP and traditional Fuzzy Logic, with AUC scores of 79.90%, 77.80%, and 71.50%, respectively, highlighting AHP’s superior accuracy and reliability for flash flood risk assessment. The results show that the southern and northeastern parts of the study area particularly Sunamganj Sadar, as well as parts of Tahirpur and Bishwambarpur Upazilas are at high risk of flash flooding. This detailed examination offers notable observations for developing targeted mitigation strategies, protecting lives and livelihoods, and Strengthening the capacity of at-risk communities and critical infrastructure.</p> Graphical Abstract <p></p>