Weeds are both a harmful pest and an important trophic resource for many biotic components. Moreover, herbicide use must be reduced to limit its impact on environment and human health. Consequently, new cropping systems are needed that both maximise weed-related biodiversity and minimise weed harmfulness. Weed dynamics models are increasingly used to design innovative cropping systems but they only consider weed densities and/or crop yield loss. Thus, the objective of the present study was to develop a set of indicators to assess weed-related harmfulness and biodiversity and to connect them to the FLORSYS model. The FLORSYS model is to date the only multispecific weed dynamics model that integrates the effect of all cropping system components (crop succession, all management techniques) in interaction with pedoclimate. In the present study, a web survey was conducted to have farmers identify and rank criteria for weed harmfulness. As a result, five harmfulness indicators were constructed: crop yield loss, technical harvest problems (green weed biomass above cutting harvest), harvest pollution (weed seed and plant biomass exported by the harvest engine), field infestation (average weed biomass in crops) and additional crop disease incidence caused by fungi-transmitting weed species (take-all disease of cereals transmitted by grass weeds). A second set of five indicators was constructed with ecologists to assess the role of weeds as a component of flora biodiversity and as a trophic resource for other biodiversity components: species richness, species equitablity, seed resource for birds during winter, seed resource for insects during spring and summer, and pollen/nectar resource for bees and other pollinators. These indicators were tested on 31 contrasting cropping systems resulting from farm surveys in two French regions. Simulations started with 16 weed species, ran over 30 years and were repeated 20 times with randomly chosen weather series from the tested region. Indicator values mostly depended on cropping system and little on year or weather. In the last step, correlations between indicators were analysed to identify synergies and antagonisms. The biodiversity indicators were all positively correlated, except that species richness was negatively correlated with species equitability and seed resources for birds or insects. The last two resources frequently co-occurred in the same situations (Spearman correlation coefficient r=0.54). Generally, equitable weed floras offered more food to birds, insects and pollinators whereas floras dominated by a small number of species were less interesting for predators. Similarly, the various harmfulness indicators were positively correlated, except additional disease risk. Harvest pollution and technical harvest problems usually occurred together (r=0.89), yield loss also increased with field infestation (r=0.59). Most of the harmfulness indicators were positively correlated with the biodiversity indicators, i.e. there was an antagonism between agricultural production and biodiversity. The correlations were though usually low (r < 0.20), indicating that many weed floras resulted in both low harmfulness and high biodiversity. Additional disease risk was not or negatively correlated to biodiversity indicators, e.g. cropping systems with low additional disease risk tended to present a more equitable weed flora (e.g. less dominated by disease-transmitting grass weeds) and to offer more lipidic seeeds for insects (e.g. less grass weed seeds). In conclusion, though there tended to be an antagonism between weed-related biodiversity and harmfulness, the tested cropping systems differed greatly and some both maximised biodiversity and minimised harmfulness. Our team is now working on identifying the characteristics of these optimal cropping systems and to propose innovative solutions for weed control and conservation.