It is important to demonstrate that replacing fossil fuel with bioenergy crops can reduce the national greenhouse gas (GHG) footprint. We compared field emissions of nitrous oxide (N2O), methane (CH4) and soil respiration rates from the C4 grass Miscanthus × giganteus and willow (salix) with emissions from annual arable crops grown for food production. The study was carried out in NE England on adjacent fields of willow, Miscanthus, wheat (Triticum aetivum) and oilseed rape (Brassica napus). N2O, CH4 fluxes and soil respiration rates were measured monthly using static chambers from June 2008 to November 2010. Net ecosystem exchange (NEE) of carbon dioxide (CO2) was measured by eddy covariance on Miscanthus from May 2008 and on willow from October 2009 until November 2010. The N2O fluxes were significantly smaller from the bioenergy crops than that of the annual crops. Average fluxes were 8 and 32 μg m−2 h−1 N2O-N from wheat and oilseed rape, and 4 and 0.2 μg m−2 h−1 N2O-N from Miscanthus and willow, respectively. Soil CH4 fluxes were negligible for all crops and soil respiration rates were similar for all crops. NEE of CO2 was larger for Miscanthus (−770 g C m−2 h−1) than willow (−602 g C m−2 h−1) in the growing season of 2010. N2O emissions from Miscanthus and willow were lower than for the wheat and oilseed rape which is most likely a result of regular fertilizer application and tillage in the annual arable cropping systems. Application of 15N-labelled fertilizer to Miscanthus and oil seed rape resulted in a fertilizer-induced increase in N2O emission in both crops. Denitrification rates (N2O + N2) were similar for soil under Miscanthus and oilseed rape. Thus, perennial bioenergy crops only emit less GHGs than annual crops when they receive no or very low rates of N fertilizer.