Tree microhabitats (cavities, conks of fungi, bark features) play an important role for both rare and specialized species biodiversity; their preservation should therefore be targeted by biodiversity-friendly forest practices. However, when compared to other old-growth characteristics like deadwood or large trees, tree microhabitats have only recently been studied so related scientific knowledge is still relatively limited. Defining target values for microhabitat densities in managed forests is mostly based on expert knowledge rather than quantitative empirical data. We compared the densities of microhabitat-bearing trees between managed forests, where wood is still harvested, and strictly protected forest reserves, where harvesting has been abandoned, in 17 French forests (222 plots) located in both lowlands and mountain regions. We found that microhabitat densities are generally higher in strict forest reserves than in managed forests and that this difference is mainly driven by standing dead and large living trees. Though scarce, standing dead trees over-contribute to the difference observed while large trees played a lesser but significant role. In addition, contrary to results obtained for other old-growth characteristics (such as deadwood volumes), the difference between managed and strict forest reserves was higher in mountain than in lowland forests. For individual microhabitats, five out of eleven microhabitats in mountains and only one in lowland (woodpecker cavities) were significantly more numerous in strict forest reserves than in managed forests. Finally, total microhabitat density and density of specific microhabitats such as cavities and bark features increased with time since the last harvest. This increase was also mainly supported by standing dead microhabitat-bearing trees. Compared to previous studies in comparable contexts, the densities we estimated were generally higher; however, such comparisons could only be made for the most documented microhabitat types. Our results support the idea that management abandonment favours the abundance and diversity of microhabitats. However, microhabitat dynamics remain poorly known and only long-term monitoring will help understand underlying mechanisms of recruitment. In the meantime, our results may inspire forest managers in their application of daily biodiversity-friendly practices.