Ship detection in Synthetic Aperture Radar (SAR) imagery is a critical yet challenging task due to the inherent noise, clutter, and complex backgrounds present in these images. In this study, we enhance the YOLO11 framework by integrating the Efficient Channel Attention (ECA) mechanism across various model sizes (YOLO11n, YOLO11s, YOLO11 m, YOLO11 l, and YOLO11x) to improve detection performance. The ECA module refines feature extraction by dynamically emphasizing salient channel-wise responses, thereby facilitating more accurate discrimination of ship features from background interference. Extensive experiments conducted on SAR datasets demonstrate that the incorporation of ECA consistently improves key performance metrics with the most pronounced enhancements observed in smaller model variants. Qualitative analyses, supported by heatmap visualizations, further confirm that the ECA-enhanced models more effectively highlight and detect ships within the complex SAR environment. Although one-way ANOVA tests did not reveal statistically significant differences at the 0.05 threshold, diagnostic assessments via Quantile-Quantile plots and residual analysis substantiate an overall positive trend in detection performance attributable to ECA integration. The results underscore the potential of attention-based mechanisms in advancing remote sensing applications, particularly in maritime surveillance, and provide a foundation for future work on optimizing lightweight detection architectures in challenging imaging conditions.