Abstarct: At the present time, with increasing industrial activities and degrading environmental issues, climate change (CC) effects as a global challenge are more obvious worldwide. Climate change can also influence on extreme events such as floods and droughts. To investigate CC at global/regional scales General Circulation Models (GCMs) are the most widely used approach to produce climate scenarios for present and future time horizons. Given the importance of GCM models as the basic tools for CC studies and with regard to various GCM models developed worldwide, selecting the best model for each study region is of considerable importance. In this study, for six regions over Iran, namely North East, North, North West, Central, South East and South West, a synoptic station was selected as representative station for that region and outputs of 25 GCM models of CMIP5 were evaluated for temperature and precipitation in the historical period (1981-2000). With regard to performance indices, the best models were selected at seasonal time-scale. These indices include Coefficient of Determination, Root Mean Square Error and Nash Sutcliffe coefficient. Finally, the outputs of selected GCM models for each region were downscaled using the proportional method for various emission scenarios (RCP2.6, RCP4.5 and RCP8.5) and for the future time period (2046-2065). The results indicated the increase in temperature in all months over all study regions. Under the RCP2.6 scenario, the lowest and highest increase in temperature was calculated under the RCP2.6 and RCP8.5 scenarios, respectively. Unlike temperature, precipitation exhibit different behavior at different areas and under various scenarios. Overall, the average total annual precipitation in the future showed increase (For example Southwest region under all scenario), decrease (For example Southeast Region under all scenario) or no change (For example Northwest under RCP4.5) compared to the historical period.