Managed forests have the potential to promote tree microhabitats and hence to conserve the biodiversity, especially of birds, bats, and saproxylic insects, associated with such features. Although the value of tree microhabitats for biodiversity has been described, surprisingly, the factors that explain the occurrence and density of tree microhabitats in managed forests remain poorly known, especially in Mediterranean forests. To address this gap, we studied the occurrence and density of nine types of tree microhabitats inventoried in 1630 trees and 59 forest stands: canopy dead wood; woodpecker cavities; non-woodpecker cavities (with a distinction between lower, medium, and upper cavities); Cerambyx cavities; loose barks or cracks; conks of fungi; and ivy. Each tree was described according to the presence/absence of microhabitats as well as tree diameter, vitality, and species. In each stand, structural variables (basal area, stem density, log volume, stand height, diameter class distribution) and the time since last cutting (i.e., the number of years since the last cutting) were assessed. At the tree level, large-diameter trees, snags, and non-coniferous species supported a higher richness of microhabitats than trees of small diameter, living trees, and conifers. Holm oak (Quercus ilex L.), which is a typical species of Mediterranean forests, exhibited a particularly high ability to host microhabitats. Tree diameter was the best predictor of the occurrence of most microhabitats (7 out of 9). We studied co-occurrences among microhabitats, but our results indicated that microhabitat types served poorly as proxies of other microhabitats due to frequent but unspecific positive relationships. At the stand level, time since last cutting was the best predictor of density of microhabitat-bearing trees and the occurrence of most tree microhabitats. The density of microhabitat-bearing trees was approximately 13 times higher in older stands (>90 years post-cutting) than in recently cut stands (90 years.