Environmental DNA is a burgeoning tool used to address wide-ranging scientific questions, including determining diets of difficult-to-sample predators. Loons are large piscivorous diving birds that capture and consume prey underwater, making it nearly impossible to visually determine their diet via observation alone. Identifying species' diets is important for understanding basic life history traits, and revealing key prey species can clarify species' roles in complex trophic webs, aid in understanding population and community dynamics, and help identify critical habitat for protection. Current information about loon diet is largely anecdotal, and traditional non-observational methods for quantifying loon diet have limitations. Analysis of eDNA from loon feces may provide biologists with a non-invasive technique for determining diet without negative sampling effects, and with increased resolution as compared to other techniques. We surveyed lakes in two areas of northern Alaska for Yellow-billed Loons (Gavia adamsii). Loon fecal samples were collected opportunistically from latrine sites without disturbing any animals and analyzed using novel marker sets to determine loon species and diet. Fish species were detected in all fecal samples, the most common being Alaska blackfish (Dallia pectoralis), and ninespine stickleback (Pungitius pungitius). This research demonstrates that eDNA metabarcoding analyses of loon fecal samples can determine the specific loon species that deposited the feces and characterize the piscine portion of their diet with limited disturbance to the animals.