Abstract: Objectives With the increasing diversification of application demands for underwater autonomous vehicles, traditional design methods centered on textual documentation have revealed numerous limitations in practical applications. To address issues such as scattered design documents, difficulty in maintenance, and low iteration efficiency among systems in conventional design methods, it is necessary to introduce a novel overall design approach. Methods By incorporating Model-Based Systems Engineering (MBSE) methodologies into the design process of underwater autonomous vehicles and integrating them with traditional design methods, a model-driven design and verification approach for underwater autonomous vehicles has been proposed. Utilizing the M-design collaborative research platform, graphical Systems Modeling Language (SysML) was employed to construct comprehensive models, including the requirement model,functional architecture model, logical architecture model and physical architecture model of the underwater autonomous vehicle, thereby establishing a complete design framework. To further validate the feasibility of the design framework, multi-system joint simulation technology was applied, and a distributed simulation platform was developed for performance simulation and verification of typical mission scenarios for underwater autonomous vehicles. Results The research findings demonstrate that model-based design methodologies can significantly enhance the design efficiency and verification capabilities of underwater autonomous vehicles, achieving a complete closed-loop process from requirements definition to design implementation. Conclusions The established design methodology provides effective guidance for the design and specification verification processes of various manned and unmanned underwater platforms.