White-footed mice prey on gypsy moth pupae while foraging for other, more abundant food. Mice appear capable of locally extirpating moths since mice exert high predation pressure on sparse pupae and are numerically decoupled from moth populations. Nevertheless, during 23 years of monitoring, moths persisted at scales .1 ha despite frequent extinctions at smaller spatial scales. We hypothesized that spatially heterogeneous intensity in mouse foraging and/or limited moth dispersal might allow moth persistence. Using a spatially explicit, individual-based, empirically parameterized model, we show that neither spatially heterogeneous predation by mice, nor limited moth dispersal alone allows moth persistence at typical mouse densities. However, synergy between both factors allows moth population persistence at naturally occurring mouse densities. For example, in models with 40 mice/ha, both limited moth dispersal with spatially homogeneous predation risk and spatially heterogeneous predation risk with unlimited moth dispersal had a 0% chance of moth persistence, but the combination of limited dispersal and heterogeneous predation risk resulted in a ~75% chance of moth persistence. Furthermore, both for limited moth dispersal with spatially homogeneous predation risk and for spatially heterogeneous predation risk with unlimited moth dispersal, moth persistence was only guaranteed at very low mouse densities, while the combination of limited moth dispersal with heterogeneous predation guaranteed moth persistence within a broad range of mouse densities. The findings illustrate a novel mechanism of ‘‘spatial selection and satiation’’ that can enhance rare species persistence under intense incidental predation by generalist predators.