Abstract: Objectives Aiming at the problem of keeping distance following of moving ships by unmanned surface vehicle, the following strategy of unmanned surface vehicle is studied. Methods Firstly, a bidirectional fitting algorithm based on polynomial and particle swarm optimization is used to eliminate the error of radar sampling data. Then, based on the motion situation of the unmanned surface vehicle and the dynamic target ship, the speed and course motion planning of the unmanned surface vehicle is carried out to keep stable distance with target ship. Results The experimental results show that the data filtered by the bidirectional fitting algorithm are smoother and conform to the ship motion law. The unmanned surface vehicle followed the target ship stably based on the planned speed and course. Conclusions This method can meet the requirements of unmanned surface vehicle following the target ship stability, and has a good effect on single target following.