Abstract: Aiming at the influence of resolution, signal to noise ratio, null and deep belief network on the performance of deep learning ship target recognition, the performance analysis of correlation recognition based on measured data is carried out. The whole experimental analysis and processing includes echo signal alignment, data selling pulse pressure processing, signal energy normalization, depth model training, classifier design and decision output. The conclusion of the experiment has laid a solid foundation for deep understanding of the principle of deep learning high resolution range profile ship target recognition technology and the engineering application design of ship target recognition.