The correct selecting of typical meteorological year is an important factor for accurate building energy simulation. In this study, the Sandia method has been applied to analyze the measured weather data of a 14-year period (1992-2005) in Bandarabass and to select the proper data for the typical meteorological year. Also, typical meteorological year has been generated by using Meteonorm and Weathergenrator softwares. Then the results of Sandia method and the two mentioned softwares have been compared with long-term average measured data for main parameters in the weather data file. Finally, those results which have the minimum difference in every month with the long-term average measured data and have best meet the persistence criterion of Hall's have been used to select the typical meteorological year. It is found that, the results of Sandia method and Meteonorm software have good agreement with the long-term average measured data during the year and have best meet with the persistence criterion of the Hall's. (C) 2009 Elsevier Ltd. All rights reserved.