With the successful completion of European Space Agency (ESA)’s PolarGAP campaign, ground gravity data are now available for both polar regions. Therefore, it is now possible to solve the GOCE polar gap problem in satellite-only gravity field recovery by using additional polar ground gravity data instead of some regularization methods. However, ground gravimetry data need to be filtered to remove the short-wavelength information beyond a certain harmonic degree to avoid spectral leakage when inferring satellite-only gravity field models. For the Arctic, the ArcGP data set was successfully applied when inferring the high-resolution gravity field model EGM2008 which could be used for this filtering there. For Antarctica, a combination of latest airborne gravimetry data from ESA’s PolarGap campaign and some previous gravity data was recently published which was irregularly distributed in space and still had some small gaps within the GOCE south polar gap. Therefore, we proposed a point mass modeling method for this filtering which was similar to the way using EGM2008 for such filtering to the ground gravity data in the Arctic. Furthermore, a variance component estimation was applied to combine the normal equations from the different sources to build a global gravity field model called IGGT_R1C. Then, this model’s accuracy was evaluated by comparison with other gravity field models in terms of difference degree amplitudes, gravity anomaly differences as well as external checking by obit adjustment and gravity data in the GOCE polar gap areas. This gravity field model performed well globally according to these checking results; especially, the RMS of the residuals between the filtered gravity data and that calculated from IGGT_R1C was the smallest (2.6 mGal in the Arctic and 5.4 mGal in Antarctica) compared with that of the relevant satellite-only gravity field models, e.g., GOCO05s. Therefore, the disturbing impact of the GOCE polar data gap problem could be solved by adding the polar ground gravity data.