Abstract: Abstract: Reference evapotranspiration (ET0) is an important component of agricultural water management, analysis on the change trend and spatial distribution characteristics of ET0 under the climate change scenarios can be used as references for estimation of crop water requirement. In the current study, daily meteorological data from 11 meteorological stations were obtained from 1961 to 2013 in hilly area of central Sichuan, Southwest China. ET0 during 1961－2013 were calculated by FAO-56 Penman-Monteith equation, ET0 during 2014－2099 were downscaled from HadCM3 (Hadley Centre Coupled Model version 3) outputs under two emission scenarios of high (A2) and low (B2) emission by statistical downscaling model (SDSM 4.2 software). Temporal trend and spatial distribution of ET0 during 1961－2099 were analyzed by Mann-Kendall trend test and Inverse Distance Weighted Interpolation method. The results indicated that ET0 decreased non-significantly (P>0.05) during baseline period (1961－2010) with a tendency rate of -5.2 mm/10a, annual mean ET0 was 757 mm, and the highest values of ET0 were found in the north and south region while the lowest values were in the central region; HadCM3 projected a statistically significant (P<0.05) and continuous increase in ET0 during 2020s (2011－2040), 2050s (2041－2050) and 2080s (2071－2099) with a rate of 8.9, 17.5 and 13.2 mm/10 a under A2 emission scenario and 12.1, 21.5 and 4.3 mm/10 a under B2 emission scenario, respectively. Averaged for the whole area, the projected increase were 7.9%, 10.9% and 16.7% under A2 emission scenario and 7.1%, 4.9% and 12.8% under B2 emission scenario for the three periods 2020s, 2050s and 2080s, respectively. The increase under A2 emission scenario was slightly higher than that under B2 scenario which may be caused by higher air temperature and solar radiation under A2 emission scenario. The highest values of ET0 were found in the northwest and south region while the lowest values were in the central region both under A2 and B2 emission scenarios for the three periods 2020 s, 2050 s and 2080 s, and the greater increase in ET0 was found in south central region while lower increase was in central, south and north region both under A2 and B2 emission scenarios for the three periods 2020 s, 2050 s and 2080 s. GCMs and SDSM were applied to project the climate changes in the future, which may lead some uncertainties in the results due to the uncertainties of GCMs, SDSMs and selection of predictors. In order to reduce the uncertainties of climate projection and improve the simulating abilities of SDSMs on future climate, more GCMs, downscaling methods and emission scenarios should be selected for the projection of ET0. This research provided possible temporal trend and spatial distribution of ET0 in the future in hilly area of central Sichuan, which can provide valuable information for irrigation schedule planning.