One of the benefits of modeling habitat selection for a given population is the ability to predict patterns in another population that inhabits an ecologically similar area. We studied habitat selection and home ranges of reintroduced and wild giant anteaters (Myrmecophaga tridactyla) in 2 South American wetlands (Iberá, Argentina, and Pantanal, Brazil). Nine reintroduced (Iberá) and 10 wild (Pantanal) adult animals were tracked via VHF and GPS between 2007 and 2015. We used resource selection functions to assess habitat selection for the wild anteaters from Pantanal. Generalized linear mixed models were constructed for resting and activity periods during both the wet and dry seasons. We then validated previous models built for reintroduced anteaters in Iberá using data from the wild animals from Pantanal. Habitat type (floodplain, grassland, open savanna, closed savanna, and forest) and distances to selected landscape traits were used as covariates. Locations near forests were positively selected in both populations. Selection of forests in Pantanal was less evident than in Iberá, probably due to the much higher availability of forests in the Brazilian site, with 38–53% of the landscape classified as good-to-high likelihood in Pantanal compared to only 4% in Iberá. Mean home-range size of males was larger in Iberá (32.50 ± 7.64 km²) than in Pantanal (14.07 ± 1.97 km²), whereas home-range sizes of females were similar in both areas (9.75 ± 1.74 km² in Iberá; 9.62 ± 2.00 km² in Pantanal). Results of this study suggest that model validation with geographically independent data is a useful tool to compare reintroduced and wild populations and to identify resources or landscape attributes that are important for a given species, even when these resources are abundant or highly available.