Mixed cropping (MC) is a key strategy to harness agriculture for climate-change. Breeding adapted genotypes can unleash the full potential of MC, both in terms of yield potential and yield stability. To achieve this goal, concepts from both breeding and ecology have to be fused in order to develop a suitable methodology for breeding for MC. In order to advance the field of breeding for MC, we evaluated yield and trait data of pure stands (PS) and mixed stands (MS) of pea (P. sativum L.) and barley (H. vulgare L.) as a legume-cereal model system. Twenty-eight pea and seven barley lines, representing European breeding material, were grown in an incomplete factorial design at two organically managed sites across two years. The general mixing ability (GMA) of pea for total mixture yield was predominant as specific mixing ability (SMA) was absent, facilitating future breeding and seed marketing efforts. The most promising pea cultivar ‘Volt’ resulted in an average total mixture yield increase of 11% (+0.43 t/ha) in MC compared to the average, while the cultivar ’Florida’ led to a yield decrease of − 31% (− 1.23 t/ha), highlighting the importance of the choice of the genotype in MS. The analysis of separated MS yields allowed to investigate the underlying mechanistic principles in genotypes’ contribution to MS yields and we revealed the major role of producer (Pr) effects in this context. The correlation between Pr effects and GMA revealed that GMA can be maximized by selecting for high Pr effects. Early vigor, onset of flowering, shoot biomass and stipule length were identified as key traits for indirect selection for high GMA in pea accounting for up to 17% of the identified variation in total mixture yield. PS yields were moderately correlated with mixture yields (r = 0.52, P = 0.013) and can serve as an additional selection criterion. Discrepancies between correlations with PS and MS yields can be exploited to identify unique MS traits that confer niche complementarity in MS. By this method we identified stipule size as such a key trait for increasing GMA of pea. Pea genotype mixtures have a stabilizing effect also in MC systems and exhibited considerably less genotype × year and genotype × location interaction than single genotypes. Our findings close existing knowledge gaps towards breeding for MC and pave the way to develop improved genotypes for diversified cropping systems as a strategy for sustainable intensification and climate change adaptation.