Abstract: As one of the leading forces in maritime battles, ships have become important detection targets in military operations. Aiming at the problems of low recognition abilities of small-scale targets in synthetic aperture radar (SAR) images of ships in large complex ocean scenes, a small target detection algorithm for ships in SAR images based on improved YOLOv8 is proposed. First, the network convolution module Conv is replaced with spatial depth non-stepped convolution to retain discriminative feature information and improve small target detection accuracy; then, based on the coordinated attention mechanism, a parallel coordinated attention mechanism is designed to strengthen the interaction in space, and perform targeted extraction and calculation of feature information in SAR images with complex backgrounds; finally, the diagonal distance measurement of the center point and the endpoint distance measurement are integrated into the distance loss function and shape loss function of SIoU respectively, and the loss function DSIoU is redefined to improve the convergence speed and detection accuracy of the network. A large number of experiments are conducted on high-resolution SAR images dataset. The results show that compared with the YOLOv8 algorithm, the accuracy of the improved algorithm is increased by 2.9 %, the number of parameters is reduced by 10.6 %, and the frame rate is increased by 13.3 %, which can meet the real-time monitoring requirements of ship targets.