Context Predators typically select prey on the basis of their availability and traits such as body size, speed, camouflage and behaviour that influence ease of capture. Such selectivity, particularly by invasive predators, can disproportionately affect the conservation status of prey. Control of top-order predators can also trigger trophic cascades if subordinate predators have different prey preference. Aims We aimed to document prey selectivity of feral cats by comparing their diet with prey availability over a 27-year study in an Australian desert. Methods Stomach-content and demographic data were recorded from 2293 feral cats, showing 3939 vertebrate prey. These were compared with vertebrate-prey availability estimated from 224,472 pitfall-trap nights, 9791 Elliott-trap nights and opportunistic sampling that accumulated 9247 small mammal and 32,053 herptile records. Potential bird availability was assessed through 2072 quantitative counts amounting to 29,832 bird records. We compared cat selectivity among species, guilds, and physical and behavioural traits of potential prey. Key results Prey guild selectivity from two quantitative subsets of these data indicated that cats preferentially selected medium-sized rodents, snakes and ground-nesting birds over other prey guilds, and also preyed extensively on rabbits, for which selectivity could not be assessed. Species that froze or responded defensively to predators were less favoured than were prey that fled, including fast-evading species. Species inhabiting dunes were hunted more frequently relative to their abundance than were closely related species on stony plains. Conclusions The size, habitat preference and response to predators of potential prey species affect their targeting by feral cats. Implications Our results assist assessment of risk to wildlife species from cat predation and suggest that cat control will trigger changes in the relative abundance of prey species depending on their size, habitat use and behaviour.