Accurate coordination of mediolateral foot placement, relative to the center of mass kinematic state, is one of the mechanisms which ensures mediolateral stability during human walking. Previously, we found that shoes constraining ankle moments decreased foot placement accuracy, presumably by impairing control over movement of the swing leg. As such, ankle moment constraints can be seen as a perturbation of foot placement. Direct mechanical perturbations of the swing leg trajectory can improve foot placement accuracy as an after-effect. Here, we asked whether constrained ankle moments could have a similar effect. If confirmed, this would offer a simple training tool for individuals with impaired foot placement control. Participants (n=19) walked in three conditions; normal (baseline, 10 minutes), while wearing shoes constraining ankle moments (training, 15 minutes), and normal again (after-effects, 10 minutes). Foot placement accuracy was calculated as the percentage of variance in foot placement that could be predicted based on the center of mass kinematic state in the preceding swing phase. When walking with constrained ankle moments, foot placement accuracy decreased initially compared to baseline, but it gradually improved over time. In the after-effect condition, foot placement accuracy was higher than during baseline, but this difference was not significant. When walking with constrained ankle moments, we observed increased step width, decreased stride time and reduced local dynamic stability. In conclusion, constraining ankle moment control deteriorates foot placement accuracy. A non-significant trend towards improved foot placement accuracy after prolonged exposure to constrained ankle moments, allows for speculation on a training potential.