Drought is the most important threat to both environmental and socio-economic features of the arid and semi-arid regions of the world as it is closely associated with agricultural production losses, water supply shortage, and security. This study aims to regionalize extreme drought severity-duration (SD) index probability across Iran. The 12-month Standardized Precipitation Index (SPI) time series were calculated for the studied stations and subsequently used to compute the annual extreme SD index. Applying a cluster analysis (CA) to the stations, SD index, the studied area was classified into four regions and their homogeneity was tested by the L-moment approach. Results indicated that the regions identified are homogeneous. Based on the L-moment ratio diagram the 2-parameter log-normal distribution was selected as the regional statistical distribution for all regions identified, by which the regional drought SD associated with different return periods was estimated. The spatial pattern of computed drought SD indicates that the eastern, northwestern and western parts of the country are exposed to very high drought severity in all return periods. As the return period is increasing, very high and severe droughts dominate larger areas, so that, for long return periods, approximately all the country would turn towards extreme and severe drought events even across regions possessing high annual precipitation amount.