Ecological theory suggests that diverse plant communities should be more resistant to plant invasion; however, empirical studies with cover crops often fail to show greater weed suppression in species-rich mixtures compared to that observed in the most weed-suppressive cover crop grown in monoculture. Our objective was to examine how the traits of cover crop species grown in mixtures influence weed suppression and to develop new metrics for assessing the relative performance of mixtures with regard to weed suppression and effects on subsequent cash crops. We conducted a two-year field experiment in southeast New Hampshire, USA. Treatments were mixtures and monocultures of annual broadleaf, legume, and grass cover crops aimed at filling three distinct temporal niches during the growing season: early-, mid-, and late-season. Cover crop and weed biomass were sampled prior to incorporation of cover crop residues. Following incorporation of residues, an indicator crop was sown to measure the effects of the previous cover crop treatment on subsequent crop performance (yield) and weed suppressiveness. We calculated the land equivalent ratio (LER) and several other synthetic indices to quantify the performance of cover crop mixtures and their effects on a subsequent test crop. For each group of cover crop treatments, cover crop mixtures tended to result in over-yielding (LER > 1), but did not suppress weeds better than the most weed suppressive cover crop grown in monoculture. Indices that accounted for individual cover crop species traits were better predictors of weed suppression and performance of a subsequent test crop than was cover crop biomass alone. Our results suggest that species trait information can be used to create cover crop mixtures that effectively enhance multiple ecosystem services, including weed suppression.