With a 2 millions of tons production, France is the second country in the European Union to produce durum wheat. Durum wheat production requires high grain nitrogen concentration. Irrigation and nitrogen fertilization must be managed simultaneously to maximize grain yield and also avoid low protein concentration and environmental impacts. To help advisors and farmers to better manage together these two agricultural operations and to develop innovative managements, developing a biodecisional model is an interesting possibility. However, knowledge is still missing on how farmers already managed these operations and how these two operations are linked. We developed the conceptual model for the decision part of this computer model. We performed a survey of 28 farmers conducted over the five French production areas investigating a diversity of growing conditions to identify the set of possible constraints and farmers’ decision rules. To analyze the survey, we first used a general inductive approach on individual cases and then built a conceptual model of the decision with a bottom-up approach. We identified four decision sequences for fertilization (N splitting, choice of N fertilizer, rate of application, fertilization triggering) and five for irrigation (irrigation period, anticipated number of irrigation cycles, irrigation cycles organization, irrigation triggering and irrigation cycle specificities). For each operation, the first three decision sequences refer to strategic decisions. The other decision sequences refer to tactical decisions. Coupling this model with a crop model could provide guidelines for managing durum wheat in the current climatic and economic changing context.