The optimization of cash crop successions and rotations is crucial for designing efficient cropping systems providing supporting ecosystem services instead of using external chemical inputs. The reintroduction of legumes is a relevant means for increasing the nitrogen resources. The interests of legumes are well known, however, they have also some weaknesses that strongly limit their adoption by farmers. One way to avoid these difficulties is to choose species and cultivars tolerant to pests and diseases and adapted to the pedoclimatic conditions. Nevertheless, when they are grown as sole crop their agronomic performance is often insufficient to obtain profitable and stable yields. Another solution is to grown legumes in intercrops with non-legume companion crops, such as cereals, in order to mitigate their poor performance. The evaluation of low input and agroecological cropping systems was carried out at INRA in the experimental station of Toulouse-Auzeville (South-West France) where 6 prototypes were compared since 2003 in 2 successive periods of 6 years. Three prototypes corresponding to 0 (control rotation: sorghum, sunflower, durum wheat), 1 (sunflower, winter pea, durum wheat) or 2 (soybean, spring pea, durum wheat) grains legumes in the 3-year rotation were compared. The same 3 rotations including multi-service cover crops (white mustard, oat/vetch, vetch or lucerne) during the fallow period between two main cash crops were also tested. Afterward a second period aiming at strongly reducing the use pesticides was carried out by redesigning 2 novel rotations based on a diversification of cultivars and species mixtures versus the control cereal-based rotation.This cropping system experiment was completed by factorial annual experimentations carried out since 2005 aiming at analyzing various types of bi-specific intercrops in order to optimize the spatial design and species and cultivars assemblages. We demonstrated that species mixtures are often effective for producing biomass in low input conditions. Moreover, intercrops are also more productive in organic farming with higher yield and greater protein content in cereals associated to grain legumes. Recently we also show that in organic farming, growing lentil with spring bread wheat was more profitable that growing lentil in sole crop. All these results shown that prototypes of agroecological cropping systems could be profitable despite a slight decrease in yields that was compensated by reduced costs. However the rotations with intercrops was less profitable under current economic conditions and European subsidy policy. The introduction of cover crops induced an increase of costs due to seed inputs and soil additional tillage, since we choose to avoid the use of glyphosate. Then we demonstrate that there is a great potential of using species diversification in space and time for improving yield and cereal grain quality in low input and organic farming.