Tropical forest restoration is becoming increasingly more applied to offset biodiversity loss and maintain ecosystem processes, but knowledge about its efficacy is still limited. We evaluated the success of tropical forest active restoration using dung beetles (Coleoptera: Scarabaeinae) as bioindicators and combining measures of species diversity, composition and functional diversity. We assessed patterns of dung beetles community assembly along a restoration chronosequence and also compared restoration areas with reference (primary and old secondary forest) and degraded (pasture) ecosystems. Species composition in the restoration areas was clearly progressing towards the preserved forests and deviating from the pasture with increasing restoration age. We also found a turnover of open environment specialists and habitat generalists to forest generalists and forest specialist species along the restoration chronosequence. However, the majority of individuals in the older restored habitats were typically forest generalists. Biomass was the only variable that increased with restoration age. Species richness, number of individuals, biomass and functional richness in the restored areas were similar to, or even smaller, than in pastures and substantially lower than forest reference sites. Rarefied richness, functional evenness and functional dispersion did not vary between the habitats. We found that while restored areas have the capacity to host forest-restricted species, 18 years since active restoration has not been long enough to recover a stable and diverse dung beetle assemblage. Our study also demonstrates that measures of composition, species diversity and functional diversity can complement each other and contribute to a better understanding of the efficacy of restoration practices.