Denitrification is an important process in the nitrogen (N) balance of intensively managed grassland, especially on poorly drained peat soils. Aim of this study was to quantify the N loss through denitrification in the top and sub soil of grassland on peat soils. Sampling took place at 2 sites with both control (0 N) and N fertilised (+ N) treatments. Main difference between the sites was the ground water level. Denitrification was measured on a weekly basis for 2 years with a soil core incubation technique using acetylene (C2H2) inhibition. Soil cores were taken from the top soil (0–20 cm depth) and the sub soil (20–40 cm depth) and incubated in containers for 24 hours. The denitrification rate was calculated from the nitrous oxide production between 4 and 24 hours of incubation. Denitrification capacities of the soils and the soil layers were also determined. The top soil was the major layer for denitrification with losses ranging from 9 to 26 kg N ha−1 yr−1 from the O N treatment. Losses from the top soil of the + N treatment ranged from 13 to 49 kg N ha−1 yr−1. The sub soil contributed, on average, 20% of the total denitrification losses from the 0–40 layer. Losses from the 0–40 cm layer were 2 times higher on the + N treatment than on the O N treatment and totalled up to 70 kg N ha−1 yr−1. Significant correlation coefficients were found between denitrification activity on the one hand, and ground water level, water filled pore space and nitrate content on the other, in the top soil but not in the sub soil. The denitrification capacity experiment showed that the availability of easily decomposable organic carbon was an important limiting factor for the denitrification activity in the sub soil of these peat soils.