Extreme rainfall estimation is an important task in engineering design projects. They generate heavy floods which can be extremely damaging and even lead to loss of lives, so the prediction of perilous rainfall is important. These predictions should be done as precisely as possible; therefore it is useful to investigate and compare the results obtained using different method. In this study, two techniques for short duration rainfall have been analysed and intensity duration-frequency (IDF) relationships developed that would help practicing hydrologist and engineers to cope-up the extreme floods. The annual maximum daily rainfall data from recent records, for the period 1985 to 2010 (Iranian calendar 1364 to 1388) were considered from two locations, i.e. Mashhad and Chabahar, situated in the east of Iran. The rainfall intensities for various return periods and annual maximum data series have been analysed using the Gumbel and the Generalized Logistic (GL) Distributions, and comparisons made between two approaches. The resulting information from Mashhad and Chabahar has been applied to extend and clarify previous intensity-duration frequency (IDF) relationships. The results show plausible agreement with other IDF curves from different countries which are presented for comparison.