Intercropping consists of growing two or more crop species simultaneously on the same field. Cereal-legume intercrops are a particularly beneficial association, with benefits in controlling the spread of diseases, insects and weeds, which can reduce chemical inputs use while improving protein content and yields. Despite all those benefits, cereal-legume intercrops are not widely grown in France as farmers face technical and economic barriers. To better support farmers in adopting cereal-legume intercropping, it is necessary to identify the factors favourable to its adoption on farms. Here, we propose a quantitative study highlighting the factors structuring the adoption of cereal-legume intercropping on French farms. Our objectives are (i) to identify the main factors fostering the adoption of cereal-legume intercropping at the national scale and (ii) to determine the existence of factors at the local level that differ from those observed at the national level. We used data from the 2020 French Agricultural Census, a national survey conducted once every ten years that provides an exhaustive photograph of the French agricultural situation (Metropolitan France and French overseas departments and regions). The survey follows two questionnaires, one more detailed than the other and only used on a sample of 70,000 farms (from a total of 450,000 farms) representative of French agriculture. We used data from the more detailed questionnaire, focusing on the 43,968 farms specialised in arable crop, mixed crop-livestock or livestock systems, which have the potential to grow mixtures of cereals and legumes. Based on a literature review, we identified 42 variables that can affect the adoption of cereal-legume intercrops. These variables cover aspects related to types of farming, organic and conservation agriculture, agroecological practices, marketing channels and involvement in groups of farmers. Then, we performed a Random Forest analysis to select the most relevant variables among those 42 at the national level. For more interpretability, we represented the results with a Classification and Regression Tree (CART). To account for local specificities, we ran the CART model obtained at the national level on three areas where we identified strong dynamics of adoption based on French agricultural statistics (Centre-East, around Jura department, dominated by dairy cattle farming; South, around Aveyron department, dominated by sheep and meat cattle farming, and Centre-West, around Loire-Atlantique department, a more diversified area, with arable, cattle, and mixed crop-livestock farming). Our results indicate that in France, organic farming, grain storage on farms, and livestock with feed autonomy for cattle and sheep are the main factors linked with the adoption of cereal-legume intercropping. Grain storage on farms and animal feed autonomy indicate that farmers use their intercrops on farms or sell them outside the agricultural cooperative channels. These findings suggest that the availability of outlets can represent a critical factor in adopting cereal-legume intercropping. Only a few agricultural cooperatives in France collect intercrops, and only in organic farming systems. The belonging to a farm machinery cooperative also came out as a factor favouring the adoption, as it can help overcome technical barriers through shared equipment and knowledge. We found the same factors at the local level but with some specificities. Our results suggest that local dynamics can be created through collective actions in the short term and the development of new market opportunities in the longer term.