As global reliance on waterway transit intensifies, it leads to a surge in vessel traffic in the sea, which has also resulted in a significant rise in maritime accidents. Investigations of these incidents identify human error as the leading cause. In order to mitigate this issue, there are growing efforts to integrate autonomous navigation technologies, aiming to reduce reliance on human decision-making. Collision avoidance and prevention play a pivotal role in ensuring maritime safety. Addressing the complex challenges of path planning in dynamic environments for autonomous vessels is critical due to their large momentum and limited maneuverability. Therefore, it is imperative to plan ship trajectories well in advance, ensuring a safe distance before any potential collision. The task of collision avoidance becomes even more challenging in restricted waterways and cluttered environments, as ships are under-actuated, lack brakes, and are difficult to control due to their high inertia. This study evaluates the performance of the Velocity Obstacle (VO) algorithm in such scenarios, as this method has already been extensively tested for ground vehicles and in some marine environments. VO aims to ensure safety by computing a ship’s velocity and course to maintain a safe separation distance. To assess its effectiveness, the method is tested in five different scenarios, including static obstacles, dynamic encounters, head-on and overtaking situations, crossing from starboard, and a cluttered environment with multiple obstacles. The analysis evaluates collision avoidance strategies in these scenarios and their ability to adhere to maritime rules, specifically the International Regulations for Preventing Collisions at Sea (COLREGs). This study aims to explore the potential of the VO approach in enhancing the safety and reliability of autonomous navigation in dynamic and constrained maritime environments.