The weed control of sunflower is a 'great challenge' for farmers throughout the World. The main goal of this study is to identify management and environmental factors which determine the weed species composition of sunflower fields. Altogether 49 sunflower fields across Hungary were surveyed for their weed flora, and 11 environmental and 19 management factors (including the use of mechanical weed control and 6 herbicide treatments) were also recorded for the same fields. Using stepwise backward selection this set of predictors was reduced to a minimal adequate model containing 14 terms explaining 37.8% of the total variation in species data. The net effects of 5 variables on species composition were significant, these were soil Mg and Ca content, preceding crop, temperature, and field size. We also performed exploratory forward/backward model selection to reveal influential predictors for several predetermined species groups and individual species. Most of the herbicides appeared to be effective against annual grass species, but no herbicide was universally effective against broad-leaved weeds. Almost all types of weeds were efficiently reduced with mechanical weed control. We obtained a relatively high share of environmental factors in the variation of species composition, which suggests that the success of agrotechnical treatments in sunflower fields strongly depends on a complex of edaphic and climatic constraints. The abundance of the most troublesome weed, Ambrosia artemisiifolia was positively correlated with high soil Ca content, lower temperature, preceding crop cereal, and lower field sizes, while it seemed to be most sensitive to fluorchloridon and propisochlor application.2