Considering the effects of an asynchronous clock and acoustic stratification,the localization problem of an underwater target node was studied when the measurement process was disrupted by unknown noise and the anchor position was uncertain.The time of flight model between underwater nodes is constructed,an interactive asynchronous communication protocol is designed,and an optimization objective function to minimize the localization error is established.An underwater target localization algorithm based on deep reinforcement learning is proposed,and layer normalization is used to improve the generalization ability of the model.Finally,simulation and experimental results validate the effectiveness of the proposed method.