Ice sheet models should be able to accurately simulate palaeo ice sheets to have confidence in their projections of future polar ice sheet mass loss and resulting global sea-level rise. This requires accurate reconstructions of the extent and flow patterns of palaeo ice sheets using real-world data. Such reconstructions can be achieved by tracing the detrital components of offshore sedimentary records back to their source areas on land. For Antarctica, however, sediment provenance data and ice sheet model results have not been directly linked, despite the complimentary information each can provide on the other. Here, we present a computational framework (Tracing Antarctic Sediment Provenance, TASP) that predicts marine geochemical sediment provenance data using the output of numerical ice sheet modelling. The ice sheet model is used to estimate the spatial pattern of erosion potential and to trace ice flow pathways. Beyond the ice sheet margin, approximations of modern detrital particle transport mechanisms using ocean reanalysis data produce a good agreement between our predictions for the modern ice sheet/ocean system and seabed surface sediments. These results show that the algorithm could be used to predict the provenance signature of past ice sheet configurations. TASP currently predicts neodymium isotope compositions using the PSUICE3D ice sheet model, but thanks to its design it could be adapted to predict other provenance indicators or use the outputs of other ice sheet models.