The agricultural research community uses languages and approaches to model farmers’ decision-making processes but does not clearly detail the steps necessary to build a decision model. We propose an original and easily applicable methodology for modelers to guide data acquisition and analysis, incorporate expert knowledge, and conceptualize decision-making processes in farming systems. It combines decision-making analysis with a modeling approach inspired by cognitive sciences and software-development methods. It is organized into four steps:1)Problem Definition Specify the context and the initial research question. Chose the type of data to collect. 2)Case Study Selection Select case studies based on theoretical sampling approach to search for diversity instead of representativeness. 3)Data Collection and Analysis of Individual Case Studies Collect preliminary knowledge on the context domain. Select appropriate knowledge representations for the task. Use critical decision method to collect farmer’s knowledge and identify critical incidents disrupting farming management. Present an initial transcription of the knowledge collected to the farmer for verification, refinement, revision. Represent farmers’ knowledge with UML diagrams.4)The Generic Conceptual Model List similarities and differences between individual analysis and UML graphs. Use literature, experts and modelers to strengthen development of the generic model by formulating and adding complex concepts. Format the generic conceptual model with UML graphs.We applied the methodology to two research questions on water management, one in adeveloped country (France) and one in an emerging country (India). This methodology can be used in different contexts and will be a useful tool to guide modelers in building decision model in farming system