Deep-sea mining has gained international interest to provide materials for the worldwide industry. European oceans and, particularly, the Portuguese Exclusive Economic Zone present a recognized number of areas with polymetallic sulphides rich in metals used in high technology developments. A large part of these resources are in the vicinity of sensitive ecosystems, where the mineral extraction can potentially damage deep-ocean life services. In this context, technological research must be intensified, towards the implementation of environmental friendly solutions that mitigate the associated impacts. To reproduce deep-sea dynamics and evaluate the effects of the mining activities, reliable numerical modelling tools should be developed. The present work highlights the usefulness of a new framework for risk and impact assessment based on oceanographic numerical models to support the adoption of good management practices for deep-sea sustainable exploitation. This tool integrates the oceanic circulation model ROMS-Agrif with the semi-Lagrangian model ICHTHYOP, allowing the representation of deep-sea dynamics and particles trajectories considering the sediments physical properties. Numerical simulations for the North Mid-Atlantic Ridge region, revealed the ability of ROMS-Agrif to simulate real deep-sea dynamics through validation with in situ data. Results showed a strong diversity in the particle residence time, with a dependency on their density and size but also on local ocean conditions and bottom topography. The highest distances are obtained for the smaller and less dense particles, although they tend to be confined by bathymetric constrains and deposited in deepest regions. This work highlights the potential of this modelling tool to forecast laden plume trajectories, allowing the definition of risk assessment scenarios for deep-sea mining activities and the implementation of sustainable exploitation plans. Furthermore, the coupling of this numerical solution with models of biota inhabiting deep-sea vent fields into ecosystem models is discussed and outlined as cost-effective tools for the management of these remote ecosystems.