Synthetic Aperture Radar (SAR) image targets usually have unobvious edge features, and target edge features at different scales are not exactly the same. Existing SAR target detection methods do not learn edge features, resulting in The model's ability to perceive edge features needs to be improved. Based on this, this paper proposes a SAR target detection method based on Edge Feature Guided Learning (EFGL). This method is based on the FCOS (Fully Convolutional One-Stage Object Detection) target detection framework and mainly uses the target edge features to guide Feature learning of Feature Pyramid Networks (FPN) explicitly enhances the network’s ability to learn edge features of targets at different scales by introducing an edge operator module into FPN; in addition, during the multi-scale feature fusion process, edges are constructed The feature-guided fusion module uses the spatial attention module that fuses edge features to achieve adjacent-level feature fusion guided by edge features. On the MSAR dataset and SAR-Aircraft-1.0 dataset, this method achieved detection accuracy of 68.68% and 67.44% respectively under the AP’07 standard, which was 1.34% and 4.81% higher than the basic network. Compared with other related algorithms, this method can perform better target positioning and has better SAR target detection performance.