The work aims to study a more effective scheduling scheme to deal with the complex waterway waters and dense logistics traffic flow in Ningbo Zhoushan Port area, to minimize the scheduling time and waiting time, that is, to maximize the efficiency. The navigation situation of Zhoushan Port in Ningbo was analyzed. The problems existing in the complex waterway area were put forward. A multi-objective function was proposed to minimize dispatching time and waiting time, a ship dispatching model for the complex waterways was established. In view of a large number of AIS data for ships, a waterway area scheduling model based on neural network was constructed, and speed change and ship prediction models were established for ships of different types and sizes to realize the prediction of ship scheduling status. An improved ship scheduling algorithm based on traditional particle swarm optimization was designed. The results showed that the model was solved by the algorithm and the traffic congestion could be judged and ship scheduling could be carried out based on different captains and spacing. Once possible congestion was predicted through the model, small vessels should pass through from the Tiaozhou Men channel and large vessels should pass through from the Xiasi Men channel and avoid congestion as much as possible. The model and algorithm can effectively improve the efficiency of ship scheduling and provide a theoretical basis for optimization of complex scheduling in shipping logistics ports.