Gravity data is modelled from the onshore Mandawa basin, a salt basin forming an isolated arm of a Mesozoic rift system situated along the sheared continental margin of East Africa. The main purpose is to verify that the Pande high, constituting the eastern rim of the Mandawa basin, contains a sedimentary sequence thick enough to allow future hydrocarbon exploration. The Bouguer anomaly response, incorporating an elevation and a terrain correction, has been used to constrain the modelling results. A regional trend is established by assuming the presence of two contrasting density bodies at a depth of 8 to 12 kilometres. Seven seismic markers have been mapped, coinciding with significant density contrasts; the shallow horizons are calibrated by the surface geology. Uncertainty on the deeper markers is high as only three deep wells penetrate the deeper sequences. The densities of the units are based on the results of these deep wells together with additional shallow hole data. The modelling results are in line with the existing seismic interpretation and hence leave scope for a sedimentary pile on the Pande basement high. The difference between the calculated model and the residual gravity field generally does not exceed 1.5 mgal. This modelling exercise illustrates that gravity data provide valuable constraints on the structural configuration of little explored basins, especially when the seismic data are of poor quality.