Marine debris is a growing problem for wildlife, and has been documented to affect more than 267 species worldwide. We investigated the prevalence of marine debris ingestion in 115 sea turtles stranded in Queensland between 2006-2011, and assessed how the ingestion rates differ between species (Eretmochelys imbricata vs. Chelonia mydas) and by turtle size class (smaller oceanic feeders vs. larger benthic feeders). Concurrently, we conducted 25 beach surveys to estimate the composition of the debris present in the marine environment. Based on this proxy measurement of debris availability, we modeled turtles' debris preferences (color and type) using a resource selection function, a method traditionally used for habitat and food selection. We found no significant difference in the overall probability of ingesting debris between the two species studied, both of which have similar life histories. Curved carapace length, however, was inversely correlated with the probability of ingesting debris; 54.5% of pelagic sized turtles had ingested debris, whereas only 25% of benthic feeding turtles were found with debris in their gastrointestinal system. Benthic and pelagic sized turtles also exhibited different selectivity ratios for debris ingestion. Benthic phase turtles had a strong selectivity for soft, clear plastic, lending support to the hypothesis that sea turtles ingest debris because it resembles natural prey items such as jellyfish. Pelagic turtles were much less selective in their feeding, though they showed a trend towards selectivity for rubber items such as balloons. Most ingested items were plastic and were positively buoyant. This study highlights the need to address increasing amounts of plastic in the marine environment, and provides evidence for the disproportionate ingestion of balloons by marine turtles.