Abstract: Typhoon storm surge disasters usually cause huge losses to coastal areas. Therefore, accurate prediction of typhoon storm surge is of practical significance for disaster prevention and mitigation in coastal areas. In this paper, a BP neural network model based on the optimization of the tunicate swarm algorithm is established based on the results of existing storm surge prediction studies, and the model is applied to the study of typhoon storm surge prediction. In this paper, 129 time-by-time data samples of typhoons affecting the Wenzhou tide station were collected and established by selecting three typhoons affecting the Wenzhou tide station as the research object. The results show that the new model overcomes the defect of BP neural network falling into local optimal solutions, and improves the convergence speed of BP neural network model based on the optimization of particle swarm optimization algorithm. The TSA-BP model performs better prediction accuracy and stability.