Urban ecosystems present ecologists with the unique opportunity to study ecological communities in the context of drastic structural and environmental change unprecedented in pristine environments. Metacommunity theory organizes a suite of predictions of how species assemble locally from the regional species pool. Understanding assemblage structure in urban ecosystems requires a revised perspective embracing human behavior and decision-making. Anthropogenic stress on the environmental certainly influences what species can colonize and thrive at a location, a process we call "passive" community assembly because humans are not purposely assembling species, but influence assembly via habitat degradation. Humans also purposely create habitat and manipulate species composition to support desirable assemblage structure, such as gardens, lawns and green space. We term this "active" community assembly. Habitat restoration is one example of such "active" assembly, and the processes underlying species assembly in restored habitats is currently being explored. Recent literature surveys and economic assessments suggest that restoration may not always result in desirable communities despite enormous economic investment. We develop here a conceptual model to help organize a revised theory of community assembly in urban ecosystems, and demonstrate how it can inform restoration science.