In the present study, the regional climate model (RegCM4) is used to project the temperature extremes of Egypt under two future scenarios (Representative Concentration Pathways: RCP4.5 and 8.5). To achieve this goal, the spatial pattern of the simulated temperature extremes is first examined. After that, the simulated temperature extremes are bias-corrected in the historical period with respect to in-situ observations using Linear-Scaling (LS) technique. Then, the LS is used to correct the two future scenarios of the period 2006-2100. The RegCM4 is downscaled by the medium resolution of the Earth System Model of the Max Planck Institute (MPI-ESM-MR) with 50 km horizontal grid spacing over Middle East and North Africa (MENA) and then eventually nested over Egypt with 20 km horizontal grid spacing. The results show that the simulated temperature extremes show a notable increase under the RCP8.5 scenario more than the RCP4.5 scenario. Furthermore, the RegCM4 performance is remarkably improved when the LS method is used. Such performance is indicated by a low mean bias in the validation period compared to the evaluation period over majority of stations. Therefore, the RegCM4 can be used to project the temperature extremes using the LS over the location of interest. In addition, using multiple General Circulation Models (GCMs) is necessary to account for the sensitivity of the RegCM4 to the atmospheric forcing.