The development of a sterilizing vaccine against malaria remains one of the highest priorities for global health research. While sporozoite vaccines targeting the pre-erythrocytic stage show great promise, it has not been possible to maintain efficacy long-term, likely due to an inability of these vaccines to maintain effector memory T cell responses in the liver. Vaccines based on human cytomegalovirus (HCMV) might overcome this limitation since vectors based on rhesus CMV (RhCMV), the homologous virus in rhesus macaques (RM), elicit and indefinitely maintain high frequency, non-exhausted effector memory T cells in extralymphoid tissues, including the liver. Moreover, RhCMV strain 68–1 elicits CD8+ T cells broadly recognizing unconventional epitopes exclusively restricted by MHC-II and MHC-E. To evaluate the potential of these unique immune responses to protect against malaria, we expressed four Plasmodium knowlesi (Pk) antigens (CSP, AMA1, SSP2/TRAP, MSP1c) in RhCMV 68–1 or in Rh189-deleted 68–1, which additionally elicits canonical MHC-Ia-restricted CD8+ T cells. Upon inoculation of RM with either of these Pk Ag expressing RhCMV vaccines, we obtained T cell responses to each of the four Pk antigens. Upon challenge with Pk sporozoites we observed a delayed appearance of blood stage parasites in vaccinated RM consistent with a 75–80% reduction of parasite release from the liver. Moreover, the Rh189-deleted RhCMV/Pk vectors elicited sterile protection in one RM. Once in the blood, parasite growth was not affected. In contrast to T cell responses induced by Pk infection, RhCMV vectors maintained sustained T cell responses to all four malaria antigens in the liver post-challenge. The delayed appearance of blood stage parasites is thus likely due to a T cell-mediated inhibition of liver stage parasite development. As such, this vaccine approach can be used to efficiently test new T cell antigens, improve current vaccines targeting the liver stage and complement vaccines targeting erythrocytic antigens.