Using the 43‐station soil moisture (SM) at a depth of 0–10 cm over eastern China, the near‐surface (about 0–5 cm) SM products from satellite missions including the level‐2 product of Chinese FengYun 3C (FY3C), level‐2 neural network product of the European Soil Moisture and Ocean Salinity (SMOS) and level‐4 product of the US Soil Moisture Active Passive (SMAP) were evaluated and compared during summers of 2015–2018. Thus, features of those products can be identified for their further application of climate study over eastern China. Large diversity is found among the satellite and field SM in terms of spatial distribution. This disagreement may cause the poor performance of the relationship between spatial coefficients of variation (CV) and mean SM. Compared with the field SM, the FY3C SM has smaller bias than the other satellite products. FY3C also performs better than SMOS in terms of the root‐mean‐square error (RMSE) and unbiased RMSE (ubRMSE) but has the smallest correlation coefficient (R). The SMAP product is generally the best among the three products in terms of RMSE, ubRMSE and R. However, a good performance in those metrics does not guarantee the same results on various time scales. On subseasonal time scales, the R in FY3C is the smallest among the three products, and the SMAP product has the largest R, but its amplitudes of the subseasonal variations are much smaller than the field observations. This indicates that when the SMAP products are applied for the analysis on subseasonal SM–atmosphere interactions, the effects of SM may be underestimated. On 10–30 days and above 60 days, dry period tends to have large spatial CV but this phenomenon is weak in all the satellite products. On the other hand, dry area tends to have large temporal CV on the four time scales, and this relationship is the strongest in SMAP but the weakest in FY3C. Therefore, there are large uncertainties of variability among different satellite SM products on subseasonal time scales over eastern China. Besides seasonal and overall performance, more attention is called to the variations on different time scales.