We evaluated the restoration of native plant assemblages by topsoil translocation in the Hunter Valley, south-east Australia. Species' responses were characterized by defining nine plant functional types (PFTs) based on combinations of four response mechanisms (seed bank persistence, germination cues, resprouting mechanisms, and longevity) through which species were predicted to persist or decline following translocation. The effects of community type and delay in topsoil restoration on restoration outcomes were tested in an orthogonal experiment. Changes in species' frequency were detected using Bayesian statistics with prior probabilities derived from pre-clearing data. Few species failed to reestablish following translocation; these were offset by recruitment of other native species not detected prior to clearing. Compositional changes were more pronounced when topsoil was stockpiled (cf direct reinstatement), although there was no trend related to the period of stockpiling. The PFT response model correctly predicted the rank probability of decline in three of the nine PFTs, while a further three were correctly placed in the top ranks but in the incorrect order. Three PFTs were incorrectly ranked because the response model was incorrect. Resprouters declined more frequently than seeders; however, species with physical seed dormancy declined less frequently than those with either transient seed banks or physiological, morphological, or morpho-physiological dormancy, irrespective of resprouting ability. Species with short juvenile periods were more likely to increase. We conclude that PFTs based on fire-response traits represent a practical means of predicting species' responses to translocation and a basis for prioritizing species for supplementary planting.