To address the challenges of low path planning efficiency for autonomous undersea vehicle(AUV) in multi-target environments, as well as the limitations of the traditional whale optimization algorithm(WOA) in terms of susceptibility to local optima and inadequate adaptability to three-dimensional obstacle avoidance requirements, this study proposed a collaborative planning strategy that integrated a fluid perturbation algorithm with an improved WOA. A hybrid population initialization method was developed by combining chaotic mapping to generate high-coverage initial solutions and a greedy algorithm to construct locally optimal sequences, effectively addressing the issue of poor solution quality caused by random initialization in traditional WOA. For the discrete characteristics of the traveling salesman problem(TSP), a discrete position update strategy based on random insertion and local inversion was proposed, significantly enhancing the algorithm’s capability to escape from local optima. An elite retention mechanism was introduced to ensure the global convergence of the algorithm through an iterative optimization framework that replaced the worst individuals with the optimal ones. During the path generation phase, a three-dimensional fluid disturbance field model was established, where obstacle perturbation matrices adjusted the original flow field direction to achieve continuous obstacle avoidance in complex obstacle environments. Simulation results demonstrate that the proposed algorithm reduces the average path length by 15.4% and 7.5% compared to traditional genetic algorithm and particle swarm optimization, respectively, while improving computational efficiency by 45.5% and 16.8%.