Abstract: Based on the observed daily precipitation data at 26 meteorological stations in the Changjiang-Huaihe River valley, and the NCEP/NCAR reanalysis data from 1961 to 1990 as well as the climate projections in middle and late 21st century by the HadCM3 model under the IPCC SRES A1B, the simulation capabilities of the statistical downscaling approach for extreme precipitation indices of the current climate have been assessed by applying canonical correlation analysis (CCA). In addition, future extreme precipitation indices in middle and late 21st century have been projected. The results show that the modeling capacity of the HadCM3 for regional climate characteristics can be effectively improved by the downscaling approach, with reducing the relative error of the climatological average state of extreme precipitation indices by 30% -100%. However, the downscaling results are still wetter in winter and dryer in summer than observations. Under the SRES A1B, frequency and intensity of extreme precipitation events are projected to increase, and the estimated increasing rate is higher for extreme precipitation index than for average one, and in summer than in winter. Fractions of extreme precipitation in middle/late 21st century are expected to increase by 14% / 25% in winter, and by 24% / 32% in summer, respectively.