<p>The parameters of the path planning algorithm for unmanned surface vehicles (USVs) usually rely on manual settings, which makes it difficult for the algorithm to achieve the optimal solution when considering multiple factors in path planning. This paper proposes a data-driven method to improve the USV path planning algorithm in response to the problem of unreasonable control caused by manual experience parameter settings in traditional algorithms. Firstly, a dataset is constructed by extracting corresponding parameters from traditional fuzzy logic control algorithms. Then, using two existing fuzzy controllers as samples, a fuzzy neural network is designed. Finally, using this dataset, a new fuzzy logic controller is generated through a fuzzy neural network. Compared with traditional fuzzy logic controllers, data-driven controllers exhibit a more reasonable distribution of variable parameters, thus verifying the superiority of neural network-based controllers. Numerical simulations show that the proposed method improves both the path length and the navigation time, while ensuring safety in complex environments.</p>