Abstract Ground-active ants were sampled from three habitats: (i) a 10-year-old Eucalyptus punctata plantation, (ii) native woodland regrowth, and (iii) the surrounding pasture, at a study site in the Hunter Valley, New South Wales, Australia. A previous study, undertaken 6 years earlier at the same study sites, revealed no difference in species richness or composition between the eucalypt plantation and pasture. The aims of the present study were: (i) to investigate the successional change in ant community structure within the plantations; and (ii) to evaluate what levels of taxonomic identification were sufficient to indicate a change had taken place. Univariate statistics (anova) were used to compare estimates of assemblage richness between habitats using data classified at five levels of taxonomic resolution: species, morphospecies, easily recognisable taxonomic units, genus and functional group. Multivariate statistics (anosim and non-metric multidimensional scaling) were used to compare ant assemblages between habitats and between sampling events at a range of taxonomic resolutions from species to functional group. This study found: (i) a significant temporal change in community composition was evident using species, genus and functional group level data, but no change was detected in the pasture or woodland; (ii) mean ant species, morphospecies and easily recognisable taxonomic units richness were significantly greater within the plantations than the pasture; (iii) compositional differences between the plantation and pasture assemblages were evident at all levels of taxonomic resolution; (iv) mean ant species and genus richness were significantly higher in the woodland than in the plantation, and these two habitats were compositionally distinct at all levels of taxonomic resolution. This is the first case study to have documented a successional response from ants to the revegetation of agricultural land with eucalypt plantations. Reasons for the temporal and interhabitat differences in community structure are discussed, as well as the implications for taxonomic sufficiency in monitoring ant community successions.