AbstractAims We aimed to investigate the effects of historical land–sea boundary and vegetation dynamics in the Australo‐Papuan region on the genetic structure of palm cockatoo populations. In doing so, we also sought to clarify the intraspecific taxonomic status of palm cockatoos, and to examine the potential conservation implications of our results.Location New Guinea and northern Australia.Methods We examined mtDNA (domain III, control region) genetic structure in 71 palm cockatoos from 17 locations across their Australo‐Papuan range.Results Twenty polymorphic sites over 242‐base pairs defined 12 haplotypes that were arranged in a 95% confidence parsimony network of six one‐step clades. Half of these were linked in one clade that included birds from eastern New Guinea–Australia, and the other half included birds from western New Guinea. Nested clade analyses revealed strong and significant genetic structure between these two clades. The average nucleotide divergence between eastern and western birds is c. 3.3%. Within the western clade there was a non‐random distribution of haplotypes according to sampling location alone, but the locations did not cluster significantly, probably due to low sample sizes. A non‐random distribution of haplotypes emerged within one of the one‐step clades from the east of the range (once rare haplotypes were removed), although the historic mechanism that may have created this pattern is unclear. The underlying low nucleotide divergence (0.39%) among haplotypes within the eastern clade suggests relatively recent common ancestry.Main conclusions Our results suggest genetic isolation of the eastern and western clades sometime during the Pleistocene. The continual reappearance of land bridges associated with Pleistocene glacio‐eustatic cycles within the eastern part of the range provides an explanation for our results. We suggest that the occurrence of two deep marine troughs maintained a narrow mountainous barrier between eastern and western birds throughout much of the Pleistocene at a time when extensive land bridges formed elsewhere in the species’ range, and that this has maintained their genetic distinctiveness. Our results provide little support for the current accepted subspecies; the western clade is roughly congruent with Probosciger aterrimus goliath (with caveats), but the otherwise unstructured small genetic distances cast considerable doubt on the remaining subspecies. The eastern and western lineages are endemic to each area and should therefore be considered for independent conservation status and management.