This paper presents a safety-certified target-reaching method for multiple underactuated autonomous surface vehicles (ASVs) operating in environments with both stationary and moving obstacles. Specifically, a nominal guidance law based on the guiding vector fields (GVFs) are employed for target-reaching, while a safety-certified guidance law based on the control barrier functions (CBFs) ensure safety. The CBFs dynamically adjust according to the type of encountered obstacle, selecting the optimal collision avoidance heading. A quadratic programming (QP) problem with guidance signal changes as the objective function and CBFs as constraints is formulated. Collision avoidance is achieved by optimizing both the ASV’s velocity and heading rate. This method guarantees that safe distances are maintained between ASVs, neighbors and obstacles while achieving target positions. Numerical simulation results substantiate the effectiveness of the proposed safety-certified control method.