Understanding the habitat requirements of forest-dependent species, such as woodpeckers, is crucial for effective management strategies in environments facing significant anthropogenic pressures. During 2013–2019, we investigated the variation in the availability of forest structural attributes in two unlogged forests, one forest with active selective logging, and one forest 33 years after it had been logged, in the Upper Paraná Atlantic Forest in Misiones, Argentina. Then, we studied the selection of the forest attributes by co-occurring populations of Helmeted Woodpecker (Celeus galeatus), Lineated Woodpecker (Dryocopus lineatus), and Robust Woodpecker (Campephilus robustus), along with their correlation to variations in home range sizes of woodpeckers. We found that the density of potential nest trees of the three woodpecker species, the density of potential Helmeted Woodpecker roost trees, the basal area ha−1 of live trees ≥ 30 cm DBH, the density of live large trees ≥ 30 cm DBH, and the volume of dead wood ≥ 30 cm diameter differed among study sites. Home range sizes of Helmeted Woodpeckers decreased with higher potential roost tree density and greater understory bamboo cover. The home range sizes of Robust Woodpeckers decreased with a higher density of dead trees, while home range sizes of Lineated Woodpeckers were not influenced by any of the variables we measured. The three woodpecker species selected diameter classes and tree species for foraging and nesting disproportionally to their availability. Helmeted Woodpecker preferred guatambú (Balfourodendron riedelianum) over other trees for roosting. We recommend considering key attributes associated with smaller home range sizes of the threatened Helmeted Woodpecker in sustainable forest management, as these attributes are also important for the other two woodpecker species. To ensure the availability of the resources selected by these woodpeckers, we recommend increasing the legal minimum cutting diameter to 60 cm for grapia (Apuleia leiocarpa), guatambú, rabo duro (Lonchocarpus leucanthus) and palo de canga (Symplocos uniflora), and to 50 cm for laurel negro (Nectandra megapotamica) and laurel amarillo (N. lanceolata). We found similar quantities of most of the tree resources we measured in unlogged forests and a forest undergoing a natural recovery for 33 years after the cessation of selective logging, and we recommend that more logged forests be restored in this manner.