This work aim to assess the “basic-quantile”and“gamma-mapping” bias correction methods in order to improve hydrological simulations of the Chiffa watershed. The results are provided by the conceptual rainfall-runoff GR2M model in the R environment, coupled with the outputs of rainfall data simulations from the RCA4 model of CORDEX-AFRICA forced by two global circulation models (MPI-ESM-LR and CRNM-CM5) under two Representative Concentration Pathways (RCP) RCP4.5 and RCP 8.5 .The validation of GR2M hydrological model on the Chiffa basin using Nash criteria made it possible to simulate future flows over the period 2074-2099. The results showed that the future flows simulated from raw rainfall data are not in adequacy with the evolution of future rainfall simulations whereas the flows simulated from the corrected rainfall data shown better results, which highlights the importance of bias correction for hydrological impact studies. Finally, the regional climate models simulate a 9% decrease in winter, and a 6% increase in spring under RCP4.5 by the end of the 21st century