With the widely used monthly gravity models, it is hard to determine the sub-monthly variations. Thanks to the high temporal resolution, a daily ITSG-Grace2018 gravity model is employed to derive the vertical deformation of the China region in 1.0° × 1.0° grids. The standard deviations of residuals between the daily and monthly averaged displacement range from 1.0 to 3.5 mm, reaching half of the median residuals, which indicates that a higher temporal resolution gravity model is quite necessary for the analysis of crustal displacement. For the signal analysis, traditional least square (LS) is limited in its analysis of signals with constant amplitude. However, geophysical signals in a geodetic time series usually fluctuate over long periods, and missing data happen. In this study, the data adaptive approach called enhanced harmonic analysis (EHA), which is based on an Independent Point (IP) scheme, is introduced to deal with these issues. To demonstrate the time-varying signals, the relative differences between EHA and LS are calculated. It illustrates that the median percentage of epochs at grids with a relative difference larger than 10% is 69.7% and the proportions for the ranges of 30%, 50%, and 70% are about 30.1%, 18.4%, and 13.0%, respectively. The obvious discrepancy suggests the advantage of EHA over LS in obtaining time-varying signals. Moreover, the spatial distribution of the discrepancy also demonstrates the regional characteristics, suggesting that the assumption of constant amplitude is not appropriate in specific regions. To further validate the effectiveness of EHA, the comprehensive analysis on the different noise types, number of IPs, missing data, and simultaneous signals are carried out. Specifically, EHA can deal with series containing white or color noise, although the stochastic model for the color noise should be modified. The signals are slightly different when selecting different numbers of IPs within a range, which could be accepted during analysis. Without interpolation, EHA performs well even with continuously missing data, which is regarded as its feature. Meanwhile, not only a single signal but also simultaneous signals can be effectively identified by EHA.