Genetic tagging, the unique identification of individuals by their DNA profile, has proven to be an effective method for research on several animal species. In this study we apply non-invasive genetic tagging from feather samples to reveal the genetic structure and estimate local population size of red-and-green macaws (Ara chloropterus) without the need to capture these animals. The study was centered in the Tambopata region of the Peruvian Amazon. Here macaws frequently visit clay licks and their naturally molted feathers provide a unique source of non-invasively sampled DNA. We analyzed 249 feathers using nine microsatellite loci and identified 221 unique genotypes. The remainder revealed 21 individuals which were ‘recaptured’ one or more times. Using a capture-mark-recapture model the average number of different individuals visiting clay licks within one breeding season was estimated to fall between 84 and 316 individuals per clay lick. Analysis of population genetic structure revealed only small genetic differences among regions and clay licks, suggesting a single red-and-green macaw genetic population. Our study confirms the utility of non-invasive genetic tagging in harsh tropical environment to obtain crucial population parameters about an abundant parrot species that is very difficult to capture in the wild.