Ship target detection in synthetic aperture radar (SAR) images is challenging due to small target sizes, low image resolution, and complex background clutter.To address these issues,this paper proposes an improved YOLOv8n-based method for SAR ship target detection.Firstly,the backbone network of YOLOv8n is replaced with CSWin Transformer to better extract useful detailed features.Secondly,a downsampling compensate module (DCM) is designed to retain more complete contextual information that might be lost during downsampling.Then,a C2f-DWR (CSPLayer_2Conv-Dilation-Wise Residual) module is proposed to enhance the network's multi-scale feature extraction capability. Furthermore,a dual dynamic token aggregator (D-Mixer) is embedded,endowing the network with stronger inductive bias and a larger effective receptive field.Finally,an improved mixed-domain attention mechanism is incorporated to improve the model′s robustness and generalization ability.Experimental results on the SSDD dataset show that compared with the baseline network, the proposed method improves mAP@0.5 and mAP@0.5∶0.95 by 1.7% and 4.2%,respectively,demonstrating significantly enhanced detection performance for SAR ship targets.