The application of building simulation for the thermal improvement of buildings in developing countries is limited by the scarcity of hourly climate data and test reference years. Therefore a generally applicable method was developed to select and combine monthly periods of hourly data, that correspond best to long-time mean values and assure representative mean temperatures and global radiation in the TRY for Central Chile. Realistic sequences of climate situations and true correlation of climate parameters result from the use of real time series of hourly data instead of a general statistical model. Standard models for missing climate variables are discussed and adapted with special emphasis on infrared sky radiation. The analysis of simulations on passive heating and cooling confirms the usefulness of this TRY for sustainable building.