The hydrodynamic performance of a vessel is highly dependent on its manoeuvring waterways. The existence of the banks and bottom, as well as the presence of the other vessels, could have a significant influence on a ship's hydrodynamic behaviour. In confined waterways, many researchers suspect the applicability of the classical potential flow method due to its nonviscous and irrotational assumption. The main objective of the present paper is to improve and develop the boundary value problem (BVP) of a potential flow method and validate its feasibility in predicting the hydrodynamic behaviour of ships advancing in confined waterways. The methodology used in the present paper is a 3D boundary element method (BEM) based on a Rankine type Green function. The numerical simulations are performed by using the in-house developed multi-body hydrodynamic interaction programme MHydro. The waves and forces (or moments) are calculated when ships are manoeuvring in shallow and narrow channels, when ships are entering locks, or when two ships are encountering or passing each other. These calculations are compared with the benchmark test data published in MASHCON (Lataire et al., 2009; Vantorre et al., 2012), as well as the published CFD (Computational Fluid Dynamics) results. It has been found that the free-surface elevation, lateral force and roll moment can be well predicted in ship-bank and ship-bottom problems. However, the potential flow solver fails to predict the sign of the yaw moment due to the cross-flow effect. When a ship is entering a lock, the return flow effect has to be considered. By adding a proper return flow velocity to the boundary value problem, the modified potential flow solver could predict the resistance and lateral forces very well. However, it fails to predict the yaw moment due to the flow separation at the lock entrance. The potential flow method is very reliable in predicting the ship-ship problem. The resistance and lateral force, as well as the yaw moment, can be predicted well by using the potential flow method.