The gradual conversion of forest stands to single-aged and single-species stands is resulting in the loss of natural roosts for many animal groups. The installation of bat boxes is one solution to compensate for the lack of natural roosting opportunities. The box models differ in their design and material, which to some extent can determine their suitability for bats. We investigated the occupancy of 187 boxes made of wood, ceramic and two sizes concrete with styrofoam; and the intra-seasonal (spring migration, breeding, mating/migration) thermal profile for each type of box. The environment of the boxes was defined by parameters that could directly affect the box’s thermal conditions (solar energy availability), or be related to sociality (distances, obstacles) or food availability (edge of forest, water). The box occupancy depended on the box type and the season: styrofoam-concrete boxes were preferred, with a higher occupancy during the mating and migration period (>75%), whereas the highest species richness occurred in ceramic boxes. Box types also differed significantly in their diurnal thermal profiles: the wooden boxes had inside temperatures similar to outside the box, whereas the styrofoam-concrete and with expanded clay aggregate (ceramic type) averaged 2°C higher. Overall, we found 6 bat species, although we concentrated our analysis on the most common species: Pipistrellus nathusii (88%). For this species, none of the parameters that could affect the box thermal condition had an impact on the occupancy, whereas the presence of obstacles and the distance to a water body, proved to be important. To ascertain that our results may be valid and be the result of differences in the box thermal properties, we tested the relationship of the box occupancy to the latitude, using data available in the literature. The occupancy of wooden boxes from the latter dataset significantly decreases with latitude, whereas for the concrete (with addition of sawdust or styrofoam) it increases, although this relationship is not significant and requires a larger sample.