Aiming at the problem that it is more difficult for radar to recognize ship targets under strong jamming, an improved model based on YOLOv5 is proposed to improve the detection accuracy and adaptability of ship targets and passive jamming. The original network is optimized by adding a detection layer for small targets in the YOLOv5 model, which aims to enhance the detection accuracy of small targets and ensure more effective recognition of target ships in complex marine environments; Due to the integration of the exponential moving average (EMA) mechanism, it can not only reduce the impact of ocean noise and complex background, but also make the model have stronger feature expression ability in recognition, thus improving the recognition performance of the algorithm. The experimental verification is carried out by using the ship RD image data set in the passive jamming environment. The experimental results show that the improved method achieves better target recognition performance on RD data sets, and effectively improves the accuracy of target recognition.