Path planning is a key technology for robots exploring the ocean floor. This paper introduces a method for robot path planning in marine environments, based on Dynamic Aggregation Strategies (DAS) and the A* algorithm. The method balances timeliness and safety to determine the optimal route for the robot. The paper first employs DAS and Density Peak Clustering (DPC) algorithms to extract path planning nodes from complex marine environments. It then uses a multi-objective weighted model to construct a regional node network. The A* algorithm is used to achieve the optimal path planning within this network. Additionally, the paper presents simulation experiments using the proposed method, demonstrating that it enables robots to navigate efficiently and safely in complex marine environments, contributing to the field of robot path planning.