First recorded in the 1970s, massive green macroalgal blooms have since become an annual recurrence in Brittany, France. Eutrophication (in particular to anthropogenic nitrogen input) has been identified as the main factor controlling Ulva ‘green tide’ events. In this study, we modelled Ulva proliferation using a two-dimensional model by coupling hydrodynamic and biological models (coined ‘MARS-Ulves’) for five sites along the Brittany coastline (La Fresnaye Bay, Saint-Brieuc Bay, Lannion Bay, Guissény Bay and Douarnenez Bay). Calibration of the biological model was mainly based on the seasonal variation of the maximum nitrogen uptake rate (VmaxN) and the half-saturation constant for nitrogen (KN) to reproduce the internal nutrient quotas measured in situ for each site. In each bay, model predictions were in agreement with observed algal coverage converted into biomass. A numerical tracking method was implemented to identify the contribution of the rivers that empty into the study bays, and scenarios of decreases in nitrate concentration in rivers were simulated. Results from numerical nitrogen tracking highlighted the main nitrogen sources of green tides and also showed that each river contributes locally to green tides. In addition, dynamic modelling showed that the nitrate concentrations in rivers must be limited to between 5 and 15 mg l- 1, depending on the bay, to reduce Ulva biomass by half on the coasts. The three-step methodology developed in this study (analysing total dissolved inorganic nitrogen flux from rivers, tracking nitrogen sources in Ulva and developing scenarios for reducing nitrogen) provides qualitative and quantitative guidelines for stakeholders to define specific nitrogen reduction targets for better environmental management of water quality.