The EPIC (Environmental Policy Integrated Climate) model was used to assess the effects of climate change on sorghum hay (Sorghum bicolor Moench.) in Southern Italy under different future climatic scenarios. The aim of this study was to compare the results of sorghum simulation obtained with a climatic baseline generated dataset with those obtained using several future datasets. The study area was located in the Capitanata plain (southern Italy). The EPIC model was calibrated and validated using experimental data sampled from a two-year experiment (2008-2009) of sorghum carried out in the experimental farm of the CRA. The baseline simulation was based on daily climatic data generated by mean of a statistical downscaling process applied on an empirical dataset including 55 years (1951-2005). The forecasting simulations data were derived by a statistical downscaling process applied on climatic projections of three general circulation models: CCSM3, ECHAM and HadCM3 and were referred to two IPCC scenarios (A2 and B1). Each general circulation models and every scenarios were run for three periods: 2011-2040, 2041-2070, 2071-2100 and so, 18 forecasting simulations were run and compared with the generated baseline. The simulation results highlight, for all three GCMs and for both scenarios, possible trends for sorghum hay in the future such as: decrease of biomass yields, crop cycle duration and irrigation volume, while an increase of irrigation water efficiency and daily evapotranspiration. These trends resulted more evident in the A2 than B1 scenario and in the last future 30-year period of simulation. With the A2 scenario a decrease up to 5 t ha-1 was obtained for dry plant biomass and a shortening up to 20 days for crop cycle duration in the third 30-year future period. CCSM3 and HadCM3 models gave results more similar to each other, different by ECHAM model that did not produce large difference between A2 and B1 climatic scenarios. EPIC model was able to simulate the response of sorghum to climate change and, based on the results obtained, some mitigation strategies could be proposed, such as: earlier sowing time or choice of cultivars with shorter growing cycle.