Combined with the underwater acoustic channel propagation loss model to select the nodes which have important locations and many neighbor nodes in all nodes as virtual beacon nodes

so that the beacon nodes can broadcast to the surrounding locations of each virtual beacon node

which can completely cover the entire network.In the process

a screening strategy is proposed to minimize the number of virtual beacon nodes. Consider the path planning of all virtual beacon nodes as the traveling salesman problem（ TSP）

and connect these virtual nodes through the Hopfield neural network to minimize the total path length. In order to solve the randomness of Hopfield neural network

and make it be applied to a large number of TSPs. In the end condition part

the idea of crossover operator Position-based Crossover is introduced

and a crossover strategy is proposed to reduce the total path length of the planning. Simulation experiments verify that this strategy can solve the path planning problem of UWSNs mobile beacon nodes

and can effectively reduce the total path length.