In the framework of the African Geoid Project, the currently available gravity data set contains many significant gaps allover the continent. A possible way to fill in these gaps before the geoid computation process, is to use a global geopotential model which best fits the African gravity field. Hence, the aim of this paper is to create a tailored reference geopotential model for Africa to be used to fill the gravity data gaps with reasonable values. With the ongoing improvement of the data base, this tailored model will also be updated iteratively. The remove-restore technique has been applied using a newly compiled \(30^{{\prime\prime}}\times 30^{{\prime\prime}}\) Digital Height Model for Africa based on SRTM. As global geopotential reference model serves EGM2008 which shows similar behaviour for Africa like GRACE/GOCE derived models. The local and global data sets, in terms of topographically-isostatically reduced gravity anomalies, are merged and used to estimate by three different techniques the potential coefficients of the tailored reference models for Africa. The used harmonic analysis techniques are the FFT technique, the least-squares technique and the Gauss-Legendre numerical integration technique. The results show that the tailored models computed within this investigation provide significantly smaller reduced anomalies (about 50 %) than those of EGM2008 or models derived from the GRACE and GOCE satellite missions for the African continent.