Retention forestry, which retains small set-asides within forests managed for timber production and other services, is an important conservation instrument for enhancing structural complexity and biodiversity in multifunctional forests. However, in contrast to local scale effects, its large-scale effectiveness is largely unknown, as this requires area-wide and sufficiently precise information on key structural elements and associated species’ habitats. Bats are particularly sensitive to forest structural characteristics and are target organisms of most retention programs. To assess their response to existing retention efforts, we here compared key habitat structures and overall habitat suitability for bats across forest areas with and without retention, using forest type and structure variables derived from remote sensing along with topographic, climatic and land-cover variables in a multi-scale modelling approach. Based on acoustic data from 135 1-hectare plots across the Black Forest, Germany, we calibrated region-wide species distribution models for 9 bat species or bat species groups thereby identifying the best-performing scale (50 – 1000 m radius) for each predictor and species(-group). Among predictors and species(-groups), forest cover and structural variables explained most (44.0 % and 38.3 %) of bat habitat selection, with forest height heterogeneity (16.4 %) and the percentage area with standing dead trees (11.7 %) performing best, mostly at small scales (50–100 m). Forests with retention showed higher values of these key structural variables, resulting in higher predicted habitat suitability for all species(-groups), highlighting positive effects of retention on structural complexity in forests and on species that benefit thereof.