In this paper, we present the knowledge-based system ROSA working on spatial and functional organizations in agriculture. The reasoning in ROSA combines hierarchical classification, case-based reasoning, and qualitative spatial reasoning. The goal of the system is twofold: formalizing and building a case base holding on farm spatial and functional organizations, and helping the analysis of new cases. Domain knowledge and cases are modeled with the help of the so-called spatial organization graphs (sogs), and represented within a description logic system. Hierarchical case-based reasoning, involving classification and qualitative spatial reasoning, is used to compare and explain farm spatial structures modeled by sogs. An example of case retrieval is proposed, followed by a global discussion on case-based reasoning in the ROSA system and related work.