Monitoring biodiversity is of increasing importance in natural ecosystems. Metabarcoding can be used as a powerful molecular tool to complement traditional biodiversity monitoring, as total environmental DNA can be analyzed from complex samples containing DNA of different origin. The aim of this research was to demonstrate the potential of pollen DNA metabarcoding using the chloroplast trnL partial gene sequencing to characterize plant biodiversity. Collecting airborne biological particles with gravimetric Tauber traps in four Natura 2000 habitats within the Natural Park of Paneveggio Pale di San Martino (Italian Alps), at three-time intervals in 1 year, metabarcoding identified 68 taxa belonging to 32 local plant families. Metabarcoding could identify with finer taxonomic resolution almost all non-rare families found by conventional light microscopy concurrently applied. However, compared to microscopy quantitative results, Poaceae, Betulaceae, and Oleaceae were found to contribute to a lesser extent to the plant biodiversity and Pinaceae were more represented. Temporal changes detected by metabarcoding matched the features of each pollen season, as defined by aerobiological studies running in parallel, and spatial heterogeneity was revealed between sites. Our results showcase that pollen metabarcoding is a promising approach in detecting plant species composition which could provide support to continuous monitoring required in Natura 2000 habitats for biodiversity conservation.