In this master thesis the flow behavior and the performance of a bow tunnel thruster at slow forward vessel motion is studied using Computational Fluid Dynamics (CFD). The study analyzes the flow and the turning ability for a cylindrical cross-section and the effect of changing the cross-section of the bow tunnel thruster. At Royal IHC it is noticed that trailing suction hopper dredgers experience a significant decrease in turning ability, using the bow tunnel thrusters, when trailing at a speed through the water of 5 [kts] in comparison to zero forward speed. Dredgers are often operating at those speeds, use the bow tunnel thruster to keep course and therefore often experience this effect in practice. To study the flow behavior the commercial CFD solver Numeca FineMarine is used. Computations are made on model scale using a simplified wedge model of a container ship (Nienhuis wedge) for validation purpose and of a trailing suction hopper dredger (Hopper wedge). For the Nienhuis wedge multiple numerical studies are performed that focus on the used set-up, non-linear iterations and convergence, time step, actuator disk modeling and first layer thickness. A grid study together with a verification and validation study close the analysis of the Nienhuis wedge. The settings from the Nienhuis wedge are used for the computations with the Hopper wedge. For the Hopper wedge a systematic tunnel cross-section variation is derived and three different shapes are computed at different ship speeds: a circular cross-section (S1010A), flattened cross-section (S0610A) and a streamlined cross-section (S0602A). The flow of the tunnel jet and the flow around the ship are comparable to the flow of a jet in a cross-flow. The flow is unsteady and fluctuates. Once the tunnel jet flow leaves the tunnel the flow interacts with the surrounding flow and is bend into the direction of that surrounding flow. A large wake region is visible behind the jet. The velocity ratio $m$ between the ship speed and the tunnel jet speed is an important factor and characterizes the behavior of the flow. At $m$=0.2 [-] a strong jet shows only little interaction with the ship flow, while a weak jet at $m$=0.4 [-] is largely influenced by the ship flow. For the Nienhuis wedge a grid study shows large numerical uncertainties. The verification and validation study shows that the computations are qualitative valid and quantitative invalid. Quantitative comparison between two different model tests shows discrepancy in the obtained side force on the wedge. However the quantitative results of this study do agree with a full scale CFD study. In both this CFD study and the full scale CFD study the hub and strut of the thruster are not modeled. It is expected that this has an effect on the side force and is a possible reason for the difference between CFD and model tests. A change in cross-section reduces the wake region behind the jet for the streamlined cross-section (S0602A) in comparison to the other cross-section. The absolute side force of the streamlined cross-section (S0602A) is significantly increased (more than 30 [\%] at $m$=0.4 [-]), while the resistance is slightly increased (4 [\%]) in comparison to the circular cross-section (S1010A) and the flattened cross-section (S0610A). The aim is an increase in absolute side force as it increases the turning ability of the ship, an increase in resistance however is negative on the fuel-consumption of the vessel. In general the cross-sectional variation shows promising results, however the numerical uncertainties of the computations are too high. It is advised to check the CFD model scale results with CFD full scale computations and to validate both with measurements. For a future CFD study it is advised to model the hub and strut of the bow thruster, because they can have an influence on the side force on the wedge.