Abstarct: In recent decades, climate change, particularly global warming has been an important issue for the international community. This problem is destructive and threatening to human life. Therefore, predicting future climate change is considered essential. One of the methods of climate change forecasting is using the air outlet of general circulation models. But considering the low spatial resolution of these models, they cannot be used for Regional Studies. To increase the power of resolution, the output of these models will be downscaled. In this research, Amameh station daily precipitation observed data in the period 1980-2005, also forecast data of 26 parameters of HadCM3 model, the fourth scenario of the International Organization for Climate Change (AR4) and 26 parameters of CanESM2 model, the fifth scenario of the International Organization for Climate Change (CMIP5) had been used. In order to downscale the output of climate change models, three methods of ANN, KNN and ANFIS were used. In this research, data of 26 parameters of two models separately as input and daily observed rainfall data were fed into the model as output. Then simulated data (Downscaled) was calculated monthly and the average of simulated data was compared with the average of monthly observed data. The result shows that KNN model is superior to the other two methods. Although all three methods of ANN, KNN, and ANFIS for both climate models of HadCM3 and CanESM2 show a good response to downscaling in this area.