Summary 1. Wildlife reintroductions can help to restore populations and save species from extinction. However, success rates of reintroductions are low due to inherent problems, such as behavioural deficiencies resulting in high post-release predation rates among captive-bred animals. In particular, the released animals may use habitats maladaptively, leading to an ecological trap, i.e. preference for low-quality habitats leading to reduced survival and/or breeding success. Ecological traps in reintroductions can be identified only through intensive studies of habitat preferences and survival of known individuals, but such studies are lacking. 2. We investigated habitat preferences and their relationship with survival by radio-tracking reintroduced, captive-bred grey partridges Perdix perdix, a widely reintroduced commercial game species and native farmland bird of conservation concern in the UK. The low success rate of grey partridge releases could be due to maladaptive habitat use and a possible ecological trap. 3. Grey partridges released as pairs in spring showed preferences for crop and field margin habitat. The use of crops had a positive, and use of field margins a negative, association with survival, suggesting field margins could serve as ecological traps for released grey partridges. Predation rates were high, and field margins probably hosted concentrated predator activity. 4. Grey partridges released as family groups (coveys) in autumn exhibited a preference for game covers, i.e. areas of tall vegetation specifically planted to provide cover for game birds, and mortality rates were lower than in spring. Habitat use did not affect survival in autumn. 5. Synthesis and applications. We have shown that newly released grey partridge pairs behave maladaptively by preferring a habitat which has a negative association with survival. In grey partridge reintroductions, we recommend releasing grey partridge family groups in autumn rather than releasing pairs of birds in spring, and providing game covers that could induce the groups to settle onto the release areas. In reintroductions in general, the habitat preferences of the released animals should be documented together with their fates, to enable detection of possible ecological traps that could threaten the success of these conservation efforts.