The AGRICORE project proposes a novel tool for improving the current capacity to model policies dealing with agriculture by taking advantage of the latest progress in modelling approaches and ICT. Specifically, the AGRICORE tool will be built as an agent-based approach where each farm is to be modelled as an autonomous decision-making entity which individually assesses its own context and makes decisions on the basis of its current situation and expectations. This modelling approach will allow simulating the interaction between farms and their context (which will account for environment, rural integration, ecosystem services, land use and markets) at various geographic scales – from regional to global. To do so, advances in big data, artificial intelligence algorithms, mathematical solvers and cloud computing services will be applied to optimise the extremely-long parameterisation and calibration phase required by current agent-based tools, to better mimic the modelling of farmers' behaviour and interactions, to credibly assess the local effects of global events and EU policies, and in general to improve policy design, impact assessments and monitoring. The AGRICORE tool will be made as a highly modular and customisable suite, and it will be released as an open-source project so institutions can transparently update and improve the tool as needs arise.