Fisheries models are generally built to assess the dynamics of a resource with a given fishing mortality which level can be decided by some “decision maker” and/or which can be an observed process, e.g. times series proportional to some observed values of fishing effort. In these cases, the resource dynamics is represented “conditionally to” fishing mortality. This mortality must be therefore primarily defined in reference to the definition of the resource Such models can be used to make stock assessment from historical data and to find which fishing mortality values could lead to some objective that can be some optimal or acceptable situation. If the values of fishing mortality can be decided by some unique institution, this knowledge of “best or acceptable mortality values” is of great interest for decision support. If not, it becomes insufficient to identify which decision of which institution could lead to obtain some given set of fishing mortalities. This situation may arise if there is a high number of fishing units belonging to various fishing fleets that may decide to generate fishing mortalities distributions among several available possibilities, according to their own interest. Examples of such problems will be presented from a Senegalese case study where several fishing fleets harvest on a common multispecies resource. Some other examples will be given from case studies identified in the context of the program ECOST with the development of a generic model which represents the joint dynamics of fishing activity and resource.