Competition for prey is thought to be important in structuring snake assemblages. However, due in part to the secretive behavior and low detectability of many snake species, this generalization is based on a limited number of studies, most of which focus on a single study site. We examined differences in diet composition, trophic niche overlap, site occupancy and detectability of five sympatric aquatic snake species between two different habitat types in the Southeastern US, replicated at the landscape scale: permanent wetlands with fishes (n = 13) and isolated, often ephemeral wetlands without fishes (n = 10). We collected >3700 prey items from snakes and compared diet composition among snake species to examine niche breadth and overlap, correcting for relative availability of prey captured independently in the same wetlands. We evaluated evidence for competitive exclusion by estimating the probability of co-occupancy for pairs of snake species in each habitat type using occupancy modeling. In wetlands with fishes, niche overlap was low, suggesting resource partitioning. Conversely, in wetlands without fishes, niche overlap was high, with most species feeding on larval or paedomorphic ambystomatid salamanders, but competitive exclusion did not occur. We suggest that high co-occupancy of aquatic snakes in wetlands without fishes despite the apparent lack of resource partitioning is due to a combination of seasonally high abundance of high quality amphibian prey, unique aspects of predator physiology and stochastic abiotic processes that prevent these systems from reaching equilibrium. Our results demonstrate that snake diets can be highly context (e.g. habitat)-specific. Studies should consider other factors in addition to competition for prey when attempting to understand snake population and community dynamics.