The False Bay Dyke Swarm represents igneous activity associated with the opening of the South Atlantic during the early Cretaceous, and can be considered a distal component of the Etendeka-Paraná Large Igneous Province. In contrast to the intense magmatism in Namibia and Brazil, with the Paraná-Etendeka continental flood basalts and huge dyke swarms, the False Bay swarm is considered to be the product of a relatively low magma-flux in a passive rift setting. Previous work suggested that the False Bay dykes consist of a single magma type, with a marked differentiation from olivine-tholeiite basalt to ferro-tholeiitic andesite, accompanied by crustal contamination. This study uses new trace element and radiogenic Sr and Nd isotope analyses to better constrain the processes of magma evolution in the dykes. Combined assimilation − crystal fractionation (AFC) modelling suggests a first stage of nearly closed system fractionation of a gabbroic assemblage (olivine + clinopyroxene + plagioclase + magnetite), accompanied by progressively larger amounts of crustal assimilation at intermediate and late stage. The AFC models show that the exposed country rocks, including Cape Granites and Malmesbury Group metasediments, are compositionally unfavorable for producing the observed assimilation trends in the dykes. Instead, a more suitable crustal assimilant would be Mesoproterozoic granitic gneisses similar to those exposed in the neighbouring Namaqua Province, which may underly the Cape Peninsula.