Red clover-grass leys incorporating three clover (Trifolium pratense L.) varieties and two grass species, tall fescue (Festuca arundinacea Schreb) and timothy (Phleum pratense L.), were sown in 2003 and grown for three harvest years (2004-2006). The crops were fertilised either once with organic fertiliser (cow manure applied in the sowing year) or yearly with mineral fertiliser. The yields of the harvested bomasses were measured and the greenhouse gas (GHG) emissions estimated by life cycle analysis (LCA) yearly for both fertilisation types. At the end of the experiment, the carry-over effect of the red clover-grass leys was studied with spring wheat (Triticum aestivum L.). Mineral fertilisation resulted in higher yields, but also about 2-3 times higher GHG emissions per unit clover-grass biomass than organic fertilisation. Grasses were favoured by mineral fertilisation, while the proportion of clover was higher and decreased less with time with organic fertilisation. Clover leys sown without grasses yielded least and produced the highest emissions per unit biomass. Tall fescue sown with clover produced more consistent yields than timothy, especially during drought stress and when the ley aged. Organic fertilisation appears the most sustainable way to produce field biomass for energy, at least when legumes are sown in the ley. After three years of clover-grass production, the biomass yield of spring wheat sown without fertiliser was about 4 t ha(-1), which could augment production of biomass per parcel, and would decrease the amount of GHG emissions from bioenergy production whether fertilised with organic or mineral fertilisers. (C) 2012 Elsevier Ltd. All rights reserved.