Inclusion of perennial vegetation filter strips (PFSs) in the toeslope of annual cropland watersheds can decrease NO₃⁻–N losses to ground and surface waters. Although PFSs are similar to riparian buffers, the processes responsible for NO₃⁻–N removal from PFSs are not well understood. Our objectives were to (i) determine the importance of denitrification as a sink for NO₃⁻–N loss from PFSs and (ii) evaluate how PFSs alter the biophysical processes that affect the relative importance of N₂O and N₂ emissions. To address our objectives, we used a coupled field laboratory approach with experimental watersheds that included the following treatments: (i) PFSs covering the bottom 10% of the watershed and an annual corn–soybean crop rotation covering the remaining upslope 90% (PFS); (ii) 100% corn–soybean rotation (CORN); and (iii) 19-yr-old 100% restored native grassland (RNG). In situ N₂O flux rates and laboratory N₂O/(N₂ + N₂O) ratios were highest in CORN watersheds followed by PFS and RNG watersheds. In contrast, potentially mineralizable C and denitrification enzyme activity (DEA) were highest in PFS and RNG watersheds and lowest in CORN watersheds. Furthermore, there was a negative correlation between N₂O/(N₂ + N₂O) ratio and DEA. In the laboratory, N₂ fluxes were highest in PFS followed by RNG and CORN. These results indicate that PFS watersheds support greater total denitrification while emitting less N₂O than croplands. Greater potentially mineralizable C in PFS and RNG suggest C availability is an important factor affecting more complete denitrification. These results suggest PFSs function similar to riparian buffers and have potential to reduce NO₃⁻–N losses from annual croplands by denitrification to N₂.