Water search and rescue (SAR) is the last line of security to protect the lives of people on the water, and the reasonable degree of the layout of the SAR site will directly affect the efficiency of SAR. For this reason, this paper proposes a SAR site layout and optimisation model based on accident black point combined with particle swarm algorithm. In the accident black spot identification stage, the equivalent number of accidents and the dynamic segmentation method are used to achieve the initial identification, followed by the introduction of the density space-based clustering algorithm (DBSCAN), which further achieves the accurate identification. Subsequently, the particle swarm algorithm (PSO) is used to find the optimal location of the SAR site with the target of the accident coverage in the water. Using the above model to validate the accident data of Shanghai Huangpu River, which is of guiding significance for the scientific layout and optimisation of the SAR sites in the Huangpu River waters.