Rising temperatures and global warming are the most important phenomena that have detrimental effects on various resources and human life. In this study, the output performance of two global models CanESM2 and HadGEM2 under three scenarios RCP2.6, RCP4.5 and RCP8.5 from LARS-WG and SDSM models were microscale to evaluate temperature changes for the period (2021-2050).The data used are the minimum and maximum temperature of Khorramabad synoptic station for the base period (1961-2010). MAE, RMSE, MSE and R2 indices were used for calibration and validation of the models and the results showed that both models have an acceptable ability in temperature simulation. The results of the forecast show an increase in the maximum and minimum temperatures under all three RCP scenarios of the HadGEM2 model in the next period compared to the base period. The highest increase in maximum temperature is related to April with 3/5% and the lowest in November is equal to 1 % and the highest change in minimum temperature is related to June with 2/4% and the lowest is related to November and October with 3/1 percent. The results of the output of the CanESM2 climate model for the next decade show that the maximum temperature, except for September, decreases by 1/2 percent in the other months compared to the base period, with the largest changes in February and March by 3/5 percent and the minimum temperature. Except for November, it will increase by 0/9 % in other andTrend analysis indicates an increase in base period temperature.