This paper wants to make the case that agent-based modelling may contribute to provide support for the difficult process of South Africa’s land reform by running scenarios that then do not need to be explored in practice. An agent-based model (ILUPSA) was developed from a database of 605 commercial farmers and 833 commercially oriented smallholders, which are the potential land redistribution beneficiaries. Three scenarios are simulated (1) when a willing buyer- willing seller mechanism (WB-WS) is used to acquire land (baseline scenario), (2) WB- WS whereas redistributed land is subdivided into viable emerging farm parcels and (3) when less productive farms are expropriated. Simulation results shows that under WB-WS only 14% of commercial farmland becomes available for redistribution. Ninety-nine percent of this land is for grazing and the remainder is field crop and horticultural land. The redistribution becomes even more marginal when only farmland with low productivity is expropriated (less than a quarter of the land that becomes available in the baseline scenario). An estimated amount of R50 billion will be required to implement land redistribution.