Feeding studies provide important information about animals and the environments in which they live. Yet little is known about the diet of the Australian Fairy Tern Sternula nereis nereis, despite the species being listed as threatened (Vulnerable) and in need of research. This study investigated the dietary composition of this bill-loading seabird, at three colony sites of an inner shelf seascape (two marine and one estuarine) using non-invasive digital photography and direct observations (number of observations [n] = 9854). Small surface schooling, inshore spawning fishes were the most important prey at all sites. Blue Sprat Spratelloides robustus, hardyheads (Atherinidae spp.), and garfishes Hyporhamphus spp. dominated the diet, contributing ≥75% of all prey at each site. The abundance of these fishes, whose spawning period overlapped the Fairy Tern breeding season in south-western Australia (October to February), is likely an important factor influencing the location of Fairy Tern colonies. Multivariate statistical analyses showed that dietary composition differed significantly among colony sites, breeding seasons, between courtship and chick feeding, and time of day. Blue Sprat, Beaked Salmon Gonorynchus greyi, and flyingfishes (Exocoetidae spp.) were present in greater proportions at Rottnest Island and Penguin Island (marine sites) than at Point Walter (estuarine). In contrast, hardyheads, Tailor Pomatomus saltatrix, and Yelloweye Mullet Aldrichetta forsteri were more common at Point Walter. Garfishes were around twice as important at Penguin Island than the other sites. Differences in habitat and fish species assemblages at each site may explain the observed spatial trends in dietary composition, while environmental factors, e.g. sea surface temperature and freshwater discharge, and natural interannual variability may explain the observed temporal trends in diet. Fish donated for courtship were ∼21% (12 mm) longer than those provisioned to chicks and the composition of prey in the diet of Fairy Terns differed between courtship and chick feeding at both Point Walter and Penguin Island. Differences in prey handling capabilities and nutritional requirements of adult females and chicks may explain these differences. Dietary composition differed significantly among diurnal periods at Point Walter and Penguin Island, with the greatest differences observed between morning and afternoon periods. At least 30 prey species were recorded, suggesting a degree of feeding opportunism. However, the large proportion of Blue Sprat, particularly at marine colony sites, highlights a potential vulnerability of Fairy Terns to changes in prey availability during their breeding period.