We describe an agent-based model that is being developed in order to simulate the spatiotemporal movement of marine mammals and boat traffic in a portion of the St. Lawrence estuary in Canada (covering the Saguenay St. Lawrence National Marine Park and the proposed St. Lawrence Estuary marine protected area). By allowing users to test the effects of different zoning scenarios for boat traffic (e.g., seasonal closures, speed limits, regulations concerning the observation of marine mammals), this model will serve as a decision support tool for the sustainable management of this unique marine ecosystem. The model may also be used to gather information about the potential effects of boat traffic on marine mammals, for example, proximity of heavy boat traffic to ecologically critical zones or the encounter rate between boats and whales related to lethal collisions. The model is being developed in Java using the Repast Simphony libraries. It includes an individual-based model of marine mammal behaviour that is being elaborated based on existing telemetry data on fin and beluga whales as well as on land based theodolite tracking of large whales. Different types of boats, including cargo ships, whale-watching boats, pleasure craft, kayaks and ferries are also modelled individually. The model attempts to reproduce the decision-making process boat captains as a function of environmental conditions and their respective goals. In this presentation, an overview of the entire system is presented, with a description of the model architecture and the approaches used to model the different types of agents.