Parasitic plants often attack multiple host species with unique defenses, physiology, and ecology. Reproductive phenology and vectors of parasitic plant genes (pollinators and dispersers) can contribute to or erode reproductive isolation of populations infecting different host species. We asked whether desert mistletoe, Phoradendron californicum (Santalaceae tribe Visceae syn. Viscaceae), differs ecologically across its dominant leguminous hosts in ways affecting reproductive isolation. Parasite flowering phenology on one host species (velvet mesquite, Prosopis velutina) differed significantly from that on four others, and phenology was not predicted by host species phenology or host individual. Comparing mistletoe populations on mesquite and another common host species (catclaw acacia, Senegalia greggii) for which genetically distinct host races are known, we tested for differences in interactions with vectors by quantifying pollinator visitation, reward production, pollen receipt, and fruit consumption. Mistletoes on mesquite produced more pollinator rewards per flower (1.86 times the nectar and 1.92 times the pollen) and received ~ 2 more pollen grains per flower than those on acacia. Mistletoes on the two host species interacted with distinct but overlapping pollinator communities, and pollinator taxa differed in visitation according to host species. Yet, mistletoes of neither host showed uniformly greater reproductive success. Fruit set (0.70) did not differ by host, and the rates of fruit ripening and removal differed in contrasting ways. Altogether, we estimate strong but asymmetric pre-zygotic isolating barriers between mistletoes on the two hosts. These host-associated differences in reproduction have implications for interactions with mutualist vectors and population genetic structure.