Farmers often grow only the few high-yielding crops in the area, and target maximum yields through high levels of inputs. These practices are based on the assumption of a positive relationship between productivity and profitability over a wide range of cropping systems (Woo et al., 2020; Erythrina et al., 2021). Part of the correlation between productivity and profitability could be related to differences in production situation and therefore yield potential, i.e. characteristics that do not depend on farmers’ decisions. For a given production situation, the systems with the highest profitability and the best environmental sustainability could be systems with moderate yields. The objective of our research is therefore to assess the relationship between productivity and profitability of cropping systems in order subsequently to help design of the most profitable and sustainable cropping systems. In this study, we use the French DEPHY network database collected on about 3,000 French commercial farms over ten years. Regression tree methods are conducted on the dataset to identify combinations of farm characteristics associated with the production situation to form a regression tree with a range of varied production situation groups. Crop yields are aggregated on the cropping system level rather than on the crop level to represent the overall cropping system productivity in each farm. Productivity is estimated by converting yields into the amount of energy produced by surface unit (GJ ha−1 yr−1), based on the energy content of each given crop calculated by the higher heating value. Profitability (€ ha−1 yr−1) is computed by the monetary value generated by the output of the cropping system. Lasso (least absolute shrinkage and selection operator) regression method (Tibshirani, 1996) is used to compute the marginal yielding effects on profitability. Treatment Frequency Index (TFI), energy use efficiency, and their interactions with productivity are also tested with their explanatory power at profitability in the Lasso regression model. Marginal yielding effects on profitability computed by the Lasso regression method is used to explore the corresponded changes in profitability resulting from a one-unit increase in productivity within each production situation group to assess the relationship between productivity and profitability of cropping systems in a consistent context.381 Under the same production potential situation, we will test the relationship between productivity and profitability notably to reveal if an intended reduction in targeted yields and an increase in the efficiency of inputs would lead to maximised overall profits compared to an increased use of inputs targeting high yields. The original oversimplified yield-profit relationship studied here under consistent production conditions from about 3,000 French commercial farms over ten years will question current advices targeting maximum yields instead of targeting maximum income and higher sustainability. Reference: Erythrina, E., Anshori, A., Bora, C. Y., Dewi, D. O., Lestari, M. S., Mustaha, M. A., ... & Syahbuddin, H. (2021). Assessing opportunities to increase yield and profit in rainfed lowland rice systems in Indonesia. Agronomy, 11(4), 777. Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society: Series B (Methodological), 58(1), 267-288. Woo, D. K., Riley, W. J., & Wu, Y. (2020). More fertilizer and impoverished roots required for improving wheat yields and profits under climate change. Field Crops Research, 249, 107756.