The development of integrated weed management strategies requires knowledge of mechanisms that influence compositional changes in weed flora. A 9‐year study was initiated in 1988 at Delhi, Canada, on a loamy sand soil to evaluate the effect of tillage systems [conventional (CT) and no‐till (NT)] and cover crops (only in NT) on weed density, species composition and associations, and crop yield in a winter wheat (Triticum aestivum L.)/bean/winter wheat rotation. Three bean types: soyabean (Glycine max L. Merr.), white bean (Phaseolus vulgaris L.) and kidney bean (P. vulgaris L.) were included. The NT system included variations: rye (Secale cereale L.) or maize (Zea mays L.) cover crop, volunteer wheat disked after harvest and wheat stubble. Data were collected in 1994, 1995 and 1996. Tillage systems, cover crops and crop type had differential effects on weed densities, species composition and associations. Weed densities were not affected by tillage or cover crops in wheat but, in the beans, densities were greater in the CT than in the NT systems. Various associations of weed species with tillage system, cover crop and crop type were observed. Crop yields were not affected by tillage type or cover crop, except that soyabean yields were highest in plots with cover crops.