The Cape south coast of South Africa has attracted significant archaeological research because it hosts the earliest evidence of human cultural and material complexity. Furthermore, the now-submerged Palaeo-Agulhas Plain provided habitat and resources for humans during the emergence of modern behavioural traits. Using a human behavioural ecology approach—optimal foraging theory—we sought to understand how this region’s flora may have contributed to the existence of early humans along the Cape south coast. We conducted monthly plant food foraging excursions over a two-year period in the seven main vegetation types that occur within the study area. Two (rarely three) local inhabitants harvested indigenous edible plant parts in 30-min foraging bouts. A total of sixty-eight participants (of Khoe-San descent), with knowledge on edible indigenous flora, contributed to 451 bouts. This study thus provides the largest published actualistic dataset on plant foraging returns. Without any prior knowledge of spatial resource density, the foragers harvested a total of 90 different edible species and obtained an overall mean (±SD) hourly return of 0.66 ± 0.45 kg/h or 141 ± 221 kcal/h. Apart from renosterveld, where winter returns were higher compared to summer, all other vegetation types showed no seasonal difference in return rates. Plants, therefore, most likely played an important role as fall-back or reliable staple food in the Cape south coast. Edible resources were unevenly distributed spatially, with calorific returns ranging from 0 to 2079 kcal/h, with fewer — but productive — high-density areas or “hotspots”. Sand fynbos (246 ± 307 kcal/h) and dune fynbos-thicket mosaic (214 ± 303 kcal/h) yielded significantly higher returns than other vegetation types (except for riparian). Prior knowledge of such hotspots, both within and amongst vegetation types, would have offered a significant foraging advantage. Finally, we provide the first quantified evidence that forager-extracted plant returns are significantly higher—nearly three times higher—in recently burnt sand and limestone fynbos vegetation. This supports Deacon’s hypothesis that hunter-gatherers could have improved their return rates by purposely burning Cape vegetation (i.e. “fire-stick farming”) to give them access to temporally abundant geophyte "hotspots". We also demonstrate the current challenges when comparing plant return rates across different studies and discuss the benefits of using participants that are not full-time hunter gatherers.