<p>In recent years, the application of multi-objective optimization algorithms to craft feasible paths considering multiple factors has garnered significant attention in handling path planning problems for autonomous underwater vehicles. However, the construction of appropriate multi-objective problem models coupled with efficient search strategies emerges as a pivotal determinant influencing the performance of multi-objective path planning algorithms. This paper introduces a multi-task assisted multi-objective optimization algorithm (MAMO) tailored to address autonomous underwater vehicle path planning problems. The proposed multi-task framework encompasses two tasks: the original path planning task and a devised simple task. These two tasks have different decision spaces due to distinct encoding strategies. Additionally, two different yet interconnected multi-objective problem models are deployed in the above two tasks. Furthermore, two knowledge transfer strategies, domain mapping-based and reconstruction-based knowledge transfer strategies, are introduced to leverage the knowledge from the simple task to assist the original task. The efficacy of the proposed MAMO is compared against eight counterparts and evaluated on three autonomous underwater vehicle path planning cases with different numbers of obstacles. The empirical findings corroborate the efficacy of the algorithm proffered.</p>