As the world's largest Roll-on/Roll-off Passenger (Ro-Pax) terminal, Zhanjiang Xuwen Port faces severe vessel scheduling pressure during peak operation periods, placing higher requirements on scheduling efficiency. To address this issue, this paper, based on previous research, proposes an optimization method for the inbound and outbound sequence of Ro-Pax vessels by further considering constraints on dedicated berth for Ro-Ro vessels carrying dangerous goods and the characteristics of variable-speed vessel navigation. First, the inbound and outbound sequencing problem of Ro-Pax vessels is modeled as an Asymmetric Traveling Salesman Problem with Time Windows (ATSPTW). Then, constraints are formulated according to the navigation rules of the harbor basin and channel, and a weight matrix reflecting the minimum scheduling time intervals between vessels was constructed. Subsequently, a bi-objective dynamic programming model is established to minimize total scheduling time and total waiting time. Finally, a bi-objective Pareto optimization algorithm extended from the Held-Karp algorithm is developed to solve the model. The numerical experiment resulst based on actual operational data show that the optimized scheduling plan significantly outperforms the actual scheduling plan in terms of both total scheduling time and total waiting time, with optimization rates reaching 6.50% and 31.09%, respectively. The simulation experiments conducted using the FlexSim 3D simulation software further verify the safety and robustness of the optimized scheduling plan, with no violation of navigation rules observed throughout the entire operation process. The proposed optimization method can significantly improve scheduling efficiency and provide decision support for vessel scheduling at Xuwen Port and other similar ports during peak operation periods.