Climate changes have a great impact on water resource availability and accessibility especially for downstream countries of transboundary rivers. Blue Nile Basin has a major contribution to the Nile basin yield. The fluctuation and changes in climate patterns in terms of precipitation in the Blue Nile Basin affect runoff and then drought and flood management strategies. The regional climate model (RCM) is a good tool for simulating and predicting the change in climate variables for the medium and long term but it has some limitations and uncertainties. Well selection of boundary conditions and parametrization of downscaling techniques for RCM will improve the prediction scenarios with limited uncertainty. This paper examines the sensitivity of changing the boundary conditions of three domains for the regional climate model “REG-CM4.3" at Blue Nile Basin with horizontal resolutions 25 * 25 km 2 for all domains. The projections outputs for all domains were investigated during the historical period from 1979 to 2005 using many statistical indicators such as Nash-Sutcliffe efficiency (NSE), Percent bias (PBIAS), cumulative distribution function (CDF), the density function of probability (PDF), the root mean square error (RMSE), observed standard deviation ratio (RSR) ratio, number of wet and rainy spells, and the precipitation concentration index (PCI). The results show that the third Domain configuration will enhance the result of the Reg-CM4.3. The outputs of this study can be used to predict future precipitation and other climate variables till 2100.