AbstractImproving techniques of flood frequency estimation at ungauged catchments is one of the major challenges for hydrologists, especially in arid and semi‐arid regions with insufficient information. Recently, popularity of flood seasonality‐based descriptors has increased among hydrologists for delineating of hydrologically homogenous regions.This study presents a data‐based comparison of three well‐known flood seasonality‐based regionalization methods in Halilrud basin in southeastern Iran. A Jack‐knife procedure is used to assess the performance of the methods for flood quantile estimation at ungauged sites. The results of these three seasonality‐based methods are compared to those results obtained from two alternative methods: a traditional regionalization method based on catchment's hydrogeomorphic characteristics similarity and a scenario that uses all available information without subdividing area.The results indicate that although peak over threshold (POT) approach which uses all flood events during year leads to better performance than other methods, but applying POT data series in a specific season (critical season) based on POT‐CS approach leads to better homogenous region and improves flood quantile estimation at ungauged sites. Copyright © 2011 John Wiley & Sons, Ltd.