Abstract: Objectives To address the problems of large parameter size, high computational cost, and difficulty in real-time deployment of synthetic aperture radar (SAR) ship detection models on resource-constrained platforms such as spaceborne and airborne systems, a lightweight SAR ship detection method was studied. Methods Using YOLOv11n as the baseline model, Partial Convolution (PConv) was first introduced to reconstruct the C3k2 module in the backbone network, so as to reduce the computational complexity during feature extraction. Then, a Lightweight Group-Attention Downsampling (LGAD) module was designed to suppress speckle noise interference through a local adaptive attention mechanism and enhance feature preservation for small ship targets. Finally, a Task-Aligned Lightweight Detection Head (TALD) was developed to alleviate the feature misalignment between classification and localization tasks by using a shared convolution architecture and a task interaction mechanism. Results Experimental results on the SSDD dataset show that the proposed method achieves an mAP@0.5 of 98.16% and a Precision of 97.26%, with about 1.2 M parameters, 3.8 GFLOPs, and an FPS of 458. Compared with YOLOv11n, the number of parameters is reduced by 53.8%, the computational cost is decreased by 39.7%, and the inference speed is improved by 12.8%. Conclusions The proposed method significantly reduces model complexity while maintaining high detection accuracy, achieves an effective balance between accuracy and lightweight design in SAR ship detection, and provides support for real-time deployment under resource-constrained conditions.