In order to accurately identify the target of the long-distance ship type under the condition of airborne photon radar, the perform plane fitting, point cloud clustering, target extraction and other processing on the scene point cloud to obtain the ship point cloud firstly. Then from the hull point cloud, three-dimensional features such as volume, surface normal vector histogram, and deck target distribution are extracted to obtain a feature array. Finally, the random forest is used to discriminate and classify the extracted features to realize the accurate identification of the hull type. Experiments show that through multiple classification experiments on 13 types of ships, the average correct recognition rate is above 95%, effectively realizing ship type recognition