Australia's vast network of inland wetlands attracts thousands of waterbirds to congregate and breed during 'boom' periods of rainfall and productivity. Monitoring waterbird biodiversity predominantly relies on annual aerial and ground-based surveys, which may miss rare or elusive species of conservation importance. Environmental DNA (eDNA) metabarcoding is being increasingly used as a tool for rapid biodiversity assessments across a wide range of ecosystems. Before eDNA can be routinely applied to new ecosystems, assays need to be optimised to address several challenges, including primer bias, reference database limitations, and data curation strategies. Here, we used an eDNA assay to detect avian species at four inland wetlands in south-eastern Australia. We assessed the impact of filtering strategies on waterbird detection probabilities and compared results against historical observational survey records. Our eDNA assay identified up to 40% of waterbirds historically recorded at the sites. Waterbirds detected with eDNA included the elusive and endangered Australasian bittern, migratory species such as gull-billed tern and whiskered tern, and an introduced feral goose not previously recorded. Community composition varied with filtering and detection approaches, highlighting the importance of assessing the impact of data curation on species detection with respect to historical species records. This study demonstrates the potential for eDNA to be incorporated into waterbird monitoring programmes. Future research should prioritise improving reference databases and optimising field sampling protocols for these ecosystems.