Restoration practitioners commonly use plantings of young seedlings to restore endangered forest communities impacted by anthropogenic disturbance, such as weed invasion and deforestation. Forest recovery may be significantly hampered by browsing pressure from herbivores, and practitioners use a variety of plant guard techniques to reduce the negative impacts of browsing on plant establishment and facilitate ecosystem regeneration. However, there is limited knowledge of which herbivore exclusion methods are most effective in protecting growing seedlings from attack and whether such benefits vary amongst different resident plant species. We compared levels of herbivore damage for 54 native rainforest tree and shrub species amongst four guard treatments that are commonly deployed by restoration practitioners to protect revegetated seedlings from attack by non-native rusa deer (Cervus timorensis) and native vertebrate herbivores during their establishment: (1) small corrugated plastic guards; (2) tall wire guards that protect single seedlings from attack; (3) small fenced exclusion plots that protect multiple revegetated seedlings as well as passively regenerating plants simultaneously; (4) unguarded plants. We also evaluated the effects of exclusion plots on the richness and composition of passively regenerating native vegetation in addition to the growth of actively revegetated seedlings through reduced herbivore activity. Herbivore activity was briefly surveyed using camera traps, and swamp wallabies (Wallabia bicolor) were found to be more active at the site than non-native rusa deer. All guard treatments significantly reduced browsing damage compared with unguarded plants, but corrugated plastic guards were less effective than wire guards and fenced exclusion plots. The fenced exclusion plots had the additional benefit of facilitating the passive regeneration of native plant species that may be cost prohibitive to regeneration otherwise. We also found that species varied in susceptibility to browsing, with eight species suffering high browse damage and five species low damage when unguarded. These findings highlight the potential for exclusion fences to provide more efficient and effective protection of browser preferred plants in the context of rainforest restoration.