With less than half of the world's tropical forests remaining, ecological restoration is urgently needed to halt biodiversity loss. However, the efficacy of different active reforestation methods remains largely untested particularly with respect to the recovery of fauna during the early years of restoration. Here, we present the results of a long‐term restoration project in the Australian Wet Tropics after 6 years of planting. Using dung beetles as bioindicators of restoration success, we investigated how the diversity and density of trees in experimental plots influence the recovery of dung beetle diversity and their ecological functions (dung removal and secondary seed dispersal). We found that after only 6 years since planting, a native dung beetle community, representing around 41% of the species found in the adjacent rainforest, has colonized the experimental plots. Plots with the highest diversity of trees (24 species planted) showed higher dung beetle diversity, dung removal, and seed dispersal but only when the density of trees on the plots was low. These plots also have higher species richness, diversity, and abundance of rainforest species, while the opposite trend was found for open‐habitat species. Therefore, planting a higher diversity of trees appears to be the best method for the early recovery of rainforest dung beetle communities and their functions. This is particularly crucial at low tree density, which is a common issue in active restoration projects as tree mortality is relatively high in the early years.