The aim of our research is to assess patterns of change in agricultural sustainability at territorial scale without losing links with the farm and food supply chain levels. The study focused on the Abondance Valley in the French Northern Alps, through a three‐strand approach: (i) surveys on farming systems and recent projects, (ii) extrapolation to all the farms recorded in the valley and building farm typology, (iii) prospective scenarios. Models and indicators were designed and used for the extrapolation phase. Three scenarios were developed ('as usual', and two 'what If' scenarios with more or less favourable local and international socioeconomic contexts), defining and mapping out patterns of change in key structural variables. Rules simulating adaptive changes were defined according to farm type and scenario. Agricultural sustainability was assessed according to these three scenarios and was characterized through indicators to grasp different factors: farming jobs, farm workload and economic viability of the farm fabric in the valley, environmental and landscape issues tied to the agro‐pastoral land uses, and the dairy processing sector issues (forage autonomy and milk volume collected). The scenario simulations highlighted the issues for label of origin sector and land use (landscape closure, manure management). In particular, the workload of the livestock farmers makes it difficult to improve grassland upkeep and preservation, whereas collective organizations would appear better placed as a solution for ensuring that all manure produced gets spread. Local stakeholders qualified the importance of the various environmental issues, and are in the process of envisaging collective actions.