Farmers facing the same production situation may have different farming objectives and thus farming practices. However, the latter are rarely precisely described since studies often use a unique indicator of soil use intensity, showing a weak explanatory value of biotic pressures. Our aim was to develop a framework to identify crop management strategies based on (i) 14 indicators of farming practices collected in 203 fields from 2004 to 2016 through interviews with the 23 farmers managing the fields, and (ii) a discussion with the farmers to assess to what extent they are valid. Then, we assessed how much these strategies helped improving our understanding of 412 weed communities surveyed in their fields. Indicators of farming intensity (e.g. treatment frequency index) was converted into ratios by dividing with the mean value of all the fields cultivated with the same crop in the same year. We identified eight crop management strategies with cluster analysis within a 950-ha area exposed to the same production situation. They differed, in decreasing order of importance, by the diversity of crops, tillage intensity, pesticide and fertiliser uses. The interviews of the farmers validated their assignment to each strategy. The interviews revealed the reasons why farmers implement a single or various strategies in their fields (seven farmers practiced two or three strategies because of differences in soil types or history of fields). The outputs of the classification gave them additional information on how their farming system differed from those of their neighbours. Using crop management strategies improved the quality of the statistical models predicting weed richness, cumulated weed richness over the 2008–2013 period, weed abundance and the frequency of weeding failure. This framework is useful to identify crop management strategies and to share information with farmers in order to be used to redesign cropping systems toward a more environmentally friendly agriculture.