The summer rainfall data in China represented by three reanalysis datasets (NCEP/DOE, ERA and JRA) are evaluated by comparing with the products of the station rainfall data in this study. Results show that the main characteritics of summer rainfall distribution in China are reasonably reproduced in three datasets, and most of the interannual and intraseasonal variations are captured. However, there are still some deficiencies in all of the three reanalysis rainfall products. In the NCEP product, an artificial rainfall center is located over the east periphery of the Tibetan Plateau, and the intraseasonal evolution of summer rainfall in midwestern and southwestern regions is poorly represented. In the ERA product, the rainfall amount is systematically underestimated and the interannual variation of rainfall over the midwestern regions is not well simulated. In the JRA product, the rainfall over coastal regions of South China is overestimated. The overestimation of weak rainfall and underestimation of heavy rainfall can be found in all of the three products. Concerning the TS and BS scores of the rainfall events with daily amount larger than 0.1 mm, the TS scrore is around 0.6 and the BS scrore is about 1.5 for all the reanalysis rainfall products. However, with the upgrade of rainfall rates, both of the scores decrease rapidly. The scores of torrential rain are very low for all reanalysis products. The analyses in this work not only indicate the capability of current advanced numerical models in simulating summer rainfall in China under the realistic circulation forcings, but also help to understand the simulation biases resulting from physical parameterization. Results would potentially contribute to the model development and evaluation in future.