Abstract: Multi-satellite remote precipitation sensing products provide a new source of data for hydrological simulation in basins with no or little observation. In this study, TRMM precipitation products (3B42V7 and 3B42RTV7) were evaluated quantitatively, and then coupled with CREST a distributed hydrological model to explore whether TRMM satellite precipitation products can substitute ground observations in the Ganjiang River basin, a typical Chinese storm region. Results indicate that 3B42 and 3B42RT fit well with ground observations, with the monthly correlation coefficients exceeding 0.9 and bias less than 5％, while the indices for daily precipitation is slightly worse. Two scenarios were also designed: Scenario I used static parameters in which gauge precipitation was used to calibrate the CREST model and satellite precipitation was used to validate it; Scenario II used dynamic parameters in which the model was recalibrated based on both satellite precipitation products. Comparison shows that the model performs better after recalibration in Scenario II, proving that mainstream satellite precipitation products have the potential to substitute gauged-based observations after model recalibration.