Understory vegetation is essential to forest functioning, integrity and provision of ecosystem services, but forest management often overlook understory conservation. As species distribution depends on many drivers, and diversity indicators varied with scales, it is difficult to define appropriate strategies for conservation, even in a single forest type. Moreover, many understory plant species occurred at both forests and non-forest ecosystems, and the assemblage could vary at sub-regional scales. Then, forest specialist species may not be similarly threatened in different ecosystem types or sub-regions. Therefore, we wonder if the variation of the understory plant assemblage of a single forest type at a sub-regional scale occurs in different forest types; how do vary the understory species assemblages compared to non-forest ecosystems at each sub-region; and how do vary the management effects on different forest types in a local scale within different sub-regions. For this, we studied understory vegetation richness and cover in three Nothofagus forest types and three non-forest ecosystems across four sub-regions of Tierra del Fuego, Argentina. We evaluated correlations, nestedness, relational patterns (by Non-metrical multidimensional scaling-NMS and Multi-response permutational procedure-MRPP) and Indicator Species Analysis-ISA, for 229 vascular plants distributed in 554 plots. Regarding local scale, we asssesed two main productive managements: timber harvesting in N. pumilio and livestock use in N. antarctica forests. Understory assemblages varied within each single forest type among different sub-regions (p < 0.033 for all MRPP), and different indicator plant species occurred at different geographic zones for two forest types (e.g., Adenocaulon chilensis in SE for N. pumilio forests, and Cotula scariosa in NW for N. antarctica forests), but not in forests with a nested understory distribution pattern (N. betuloides). Ecosystem types shared a high quantity of species (>40 %), except in SE zone. However, all ecosystem types significantly differed in every zone (p < 0.025), except grasslands and shrublands in NW (p = 0.165). Grasslands stand out as the species richest ecosystem in NW, SW and SE zones (between 62 and 123 spp.), highlighting their importance for conservation. Landscape heterogeneity emerged as an important mechanisms enriching species richness from neighbouring habitats, while in managed stands, resource availability seemed the main driver. Regarding the local scale, higher impacted by harvested condition in N. pumilio forests showed differences with less impacted conditions despite the zone, with some common indicator species (e.g., Taraxacum officinale). However, differences between aggregate retention and primary forests depends on the locality. Concerning N. antarctica forests, livestock homogeneizes the assemblages between zones, although indicator species remains different across zones, with dominance of exotics introduced for forage in NE (e.g., Poa pratensis). The presence of the same species in forests and non-forest ecosystems reduce the local extinction hazards in Tierra del Fuego. However, conservation strategies should be implemented at every sub-region. New natural reserves could be an alternative for conservation in zones without natural reserves, but complemented with land-sharing in managed forests. The consideration of potential differences among geographic zones could bring a more clear alternative to improve conservation at sub-regional scales where the same forest type cover large regions.