Understanding the factors, or filters, that preclude the recolonization of, and population persistence in, restoration areas by fauna is important for maximizing the biodiversity benefits of restoration. We investigated filters through a long-term study examining reptile and mammal recolonization of, and population persistence in, post-mining restoration in the northern jarrah forest of southwestern Australia. Neither mammal community composition, nor individual mammal species, differed between unmined forest and restoration but reptile communities did and some reptile species were significantly less abundant in restoration than in unmined forest. We concluded that there were no filters to mammal recolonization and persistence but that there were both unidirectional (decreasing gradually over time) and dynamic (fluctuating over time) filters to reptile recolonization of post-mining restoration that influenced both recolonization and population persistence of some reptile species. Low volumes of coarse woody debris appeared to be a unidirectional filter that will slow recolonization by some reptile species for decades or even centuries. Canopy cover appeared to be a dynamic filter that fluctuated over a period of >7 years and precluded population persistence of some reptile species that had already recolonized the post-mining restoration. A dearth of studies made it difficult to generalize how widespread these filters were in restoration areas but studies of thinning and burning in natural forests suggested they were likely to be widespread. Although management techniques to reduce the effect of unidirectional and dynamic filters remain poorly developed, their development and adoption will improve the ability of restoration to facilitate fauna recolonization and persistence and better contribute towards biodiversity conservation.