Maintaining yield when reducing inputs is one prime objective of sustainable agriculture. In this context, cereal-legume intercropping is a practice that can achieve increased yield under low-input conditions through the complementary use of abiotic resources and facilitation mechanisms. Many management options exist to design cereal-legume intercropping systems, among which the choice of the species intercropped and the level of nitrogen (N) fertilization are essential. In this study, we collected the results of 35 field experiments across Europe of cereal-grain legume intercrops that combined various intercropped species and N fertilization levels. We first assessed the intensity of the biodiversity effect and its components in unfertilized intercrops. Then, we focused on a subset of systems to analyze how N fertilization influenced biodiversity effects on three intercrops (durum wheat/pea, soft wheat/pea, and durum wheat/faba bean). The biodiversity effect represents the gap between the observed and expected yields of a mixture. The complementarity effect is the performance of mixtures relative to the performance of the component monocultures. The selection effect captures the extent to which a species with a high monoculture yield dominates a mixture at the expense of the other intercropped species. Our results confirmed an overall positive biodiversity effect under unfertilized conditions and various climate conditions (0.86 ± 0.04 t.ha−1). Complementarity effect was the main driver as it represented 76% of the biodiversity effect, confirming intercropping as a useful practice in low-input systems. N fertilization lowered the complementarity effect in durum wheat/pea intercrops, did not influence these effects in soft wheat/pea intercrops, and increased only the selection effect in durum wheat/faba bean intercrops. These results highlight the need for a sufficiently competitive legume in intercrops when N fertilizers are applied in order to avoid too much disruption of plant–plant interactions.