Abstract: Accurate and continuous precipitation is vital for water resource management, flood forecasting, and hydrologic process simulation, although traditional gauge-based precipitation data is often insufficient due to the sparse and uneven distributions of stations. Moreover, it's always difficult to improve the precision and performance by interpolation technique. Satellite-based precipitation product solves these deficiencies of gauge-based data and provides an alternative data source for ungauged regions due to its high spatial and temporal resolution. The following research. This study mainly quantitatively evaluates the accuracy and suitability of the latest version (V7) of TRMM (Tropical Rainfall Measuring Mission) 3B42 multi-satellite precipitation estimation product in Pearl River basin, China, and to perform a hydrological simulation to further verify the applicability of the data. Dongjiang River basin and Beijiang River basin, located in the downstream area of Pearl River basin, were selected as the study case. Statistical indices of the Dongjiang River basin and the Beijiang River basin, including correlation coefficient (R), root mean square error (E_RMS), mean error (E_M), mean absolute error (E_MA), relative error (E_R) and Nash Sutcliffe Coefficient Efficiency (E_NSC) were utilized to measure the accuracy of 3B42-V7 data. By contrasting the quantitative assessment results of 3B42-V7 data and gauge precipitation data derived from 90 rainfall stations during 1998-2006, the R values (correlation coefficients) of majority of the grid cells were greater than 0.60 on a daily scale and greater than 0.90 on a monthly scale, suggesting that the 3B42-V7 data presents high accuracy in the studied area. Moreover, E_M (at two time scale) and E_R are higher than 0 in most grid cells, which imply that a systematic overestimation of the precipitation in the studied area exist in the 3B42-V7 data. In the moderate heavy rainfall class, the results of intensity distribution of 3B42-V7 data were similar to that of gauge data, while 3B42-V7 data underestimated the small rainfall class while overestimating the heavy rainfall class, indicating that the 3B42-V7 data had relative low accuracy in detection of extreme precipitation events. Spatial distribution of the statistical indices of 3B42-V7 data implies that the accuracy of satellite-based precipitation data is typically affected by altitude. VIC (Variable Infiltration Capacity) model, a large-scale distribution hydrological model, was applied to hydrological simulation in this case. Two scenarios are set in this study: scenario I, 3B42-V7 data were applied to the model calibrated with gauge data while they showed a relatively unsatisfactory performance during the hydrological simulation; scenario Ⅱ, 3B42-V7 data performed preferably with the model recalibrated by the 3B42-V7 data, suggesting that calibrating the model with 3B42-V7 data rather than gauge data before the hydrological simulation would be ideal. The results of the hydrological simulation validation implied that, to some extent, the 3B42-V7 products can be used as the data source in the ungauged regions.