In Europe, many water resources systems are in structural water deficit, with limited or non-existent opportunities to store more water. Often, the agricultural sector is blamed for taking water when it is least available. This paper describes a fine, spatially and temporally explicit agricultural model which was built specifically to support the design and the assessment of alternative cropping system spatial distributions to face water scarcity issues at river basin level. We focus on the description of decision making process and spatio-temporal organization of irrigation at farm level. Information available in the French “Land Parcel Identification System” (LPIS) geographical database was crucial to the modeling process. The agricultural model is coupled to a hydrological and a normative model in the MAELIA multi-agent simulation platform which aims at representing of the structure and dynamics of a particular social-ecological system: a socio-agro-hydro-system. We give an overall description of the model components and present the main simulation results used to assess the models’ performance. We discuss the model’s ability to figure inter and intra annual irrigation withdrawal dynamics.