Three measurement campaigns were carried out to answer questions related to the factors controlling variations in nitrous oxide (N2O) fluxes from intensively managed grassland on peat soil, comparison of flux measurements with a closed flux chamber method and a flux gradient technique and the development and testing of a simple empirical model for the estimation of N2O fluxes from intensively managed grassland on peat soils. Fluxes of N2O were measured with 42-48 flux chambers and ranged from less than 0.01 to 6.66 mg N/msuperscript 2 per hr. Fluxes were significantly correlated with denitrification activity (Rsuperscript 2=0.34-0.56). Contents of nitrate (NO3-) and ammonium (NH4 ) in the top soil and the water-filled pore space (WFPS) explained 37-77% of the variance in N2O flux. Spatial variability of N2O fluxes was large with coefficients of variation ranging from 101 to 320%. Spatial variability was suggested to be related to distribution of mineral N fertilizer and cattle slurry, urine and dung patches and variations in groundwater level within the field. Average field fluxes obtained with the closed flux chamber method were about a factor 10 larger than those with the flux gradient technique on one measurement day but were similar on two other measurement days. The results of the measurement campaigns were used to derive a simple empirical model including total mineral N content and WFPS. This model was tested using an independent data set, i.e. the results of a monitoring study of two years carried out on two other grassland sites on peat soil. The model reasonably predicted magnitude of and temporal variations in N2O fluxes. It is suggested that a simple empirical model which requires only easily obtainable data such as mineral N content and moisture content, in combination with a few days lasting measurement campaigns, may be a valuable tool to predict N2O fluxes from similar sites.