The adaptability of recent ocean tidal models and Earth tidal models is investigated comprehensively by means of 22 high precision tidal gravity observation series at 20 stations of the Global Geodynamics Project. Careful preprocessing of the original observations was carried out using international standard algorithms and the tidal gravity parameters were computed. The gravity load vectors of 8 main constituents are obtained based on loading computation theory and various global ocean models. The loading corrections of 14 secondary constituents are obtained based on a two-dimensional interpolation technique. Considering different characteristics of the wave amplitude, a method of “non-identical weighted mean” is developed for computing the av-eraged observed residual and remaining residual vectors at each station. The efficiency of the loading correction and the discrepancy between corrected amplitude factors and theoretical ones are analyzed. Meanwhile the calibration problem of the instruments is also discussed. After loading correction, the averaged tidal gravity parameters for all stations are obtained. The results show that the discrepancies between the global mean amplitude factors and theoretical values are less than 0.3%, the largest calibration error of the instruments is less than 0.5%. On the other hand, there are indications that the slight phase advance of K1 with respect to O1 in Mathews’ theory could be verified by ground based tidal gravity observations. The adaptability of recent ocean tidal models and Earth tidal models is investigated comprehensively by means of 22 high precision tidal gravity observation series at 20 stations of the Global Geodynamics Project. Careful preprocessing of the original observations was carried out using international standard algorithms and the tidal gravity parameters were computed. The gravity load vectors of 8 main constituents are obtained based on loading computation theory and various global ocean models. The loading corrections of 14 secondary constituents are obtained based on a two-dimensional interpolation technique. Considering different characteristics of the wave amplitude , a method of “non-identical weighted mean” is developed for computing the av-eraged observed residual and residual residual vectors at each station. The efficiency of the loading correction and the discrepancy between corrected amplitude factors and the theoretical ones are analyzed. Meanwhile the calibration problem of the results is that the discrepancies between the global mean amplitude factors and the theoretical values ​​are less than 0.3%, the largest calibration error of the instruments is less than 0.5%. On the other hand, there are indications that the slight phase advance of K1 with respect to O1 in Mathews’ theory could be verified by ground based tidal gravity observations.