To gain information about the possible use of legume cover crops as an alternative and sustainable weed-control strategy for winter wheat (Triticum aestivum L.), an experiment was conducted at two sites in the Swiss Midlands in 2001/2002. Under organic farming conditions winter wheat was direct-drilled into living mulches established with four different legume genotypes or into control plots without cover crops. Compared to NAT (control plots without cover crops but with a naturally establishing weed community), white clover (Trifolium repens L.), subclover (Trifolium subterraneum L.), and birdsfoot trefoil (Lotus corniculatus L.) reduced the density of monocotyledonous, dicotyledonous, spring-germinating, and annual weeds by the time of wheat anthesis. Strong-spined medick (Medicago truncatula Gaertner) was less efficient in this regard. While the grain yield was reduced by 60% or more for all legumes when compared to NOWEED (control plots kept weed-free), a significant negative correlation between the dry matter of the cover crop and weeds as well as between the cover crop and the winter wheat was observed by the time of wheat anthesis. The effect of manuring (60m 3 ha −1 liquid farmyard manure) was marginal for weeds and cover crops but the additional nutrients significantly increased total winter wheat dry matter and grain yields. The suppression achieved by some legumes clearly demonstrates their potential for the control of weeds in such cropping systems. However, before living legume cover crops can be considered a viable alternative for integrated weed management under organic farming conditions, management strategies need to be identified which maximise the positive effect in terms of weed control at the same time as they minimise the negative impact on growth and yield of winter wheat.