The proposed methodology, MaSID, brings an agile approach to the development of intelligent decision support systems, in this sense the work aims to provide a guide for planning and development in order to correlate external information, objective or subjective, that comes to reinforce or modify a decision, thus modifying behaviors internal to the system. Current agile methodologies do not include requirements gathering nor do they lead to an analysis of how external factors can influence decisions. To fill this gap, our methodology has an iterative and incremental flow where each phase is made up of three 15-minute work meetings each, which receive input artifacts and generate output artifacts, the result of the analyzes performed. A cycle of the methodology represents a passage through all phases, while the execution of a phase means that all planned meetings have been carried out. The conduct of meetings aims to ensure greater efficiency, collaboration and adaptation to changes. For deeper analyzes during meetings, the methodology provides conceptual frameworks and specific artifacts. We can mention some such as, the negotiation framework, artifacts in the format of tables and forms and the 3D complexity graph that allows you to visualize the size and complexity of the desired software. These tools enable a collaborative environment during the meeting, and in addition, they improve the cognitive process of analysis, bringing a continuous feedback cycle. Therefore, we understand that the MaSID methodology can improve the quality and objectivity of intelligence criteria definitions in a decision support system. Part of the methodology was validated in a project focused on gasoline import contracts, exposed in the first case study.