Aiming at the difficult recognition problem caused by the different sizes of aircraft detection and complex backgrounds in remote sensing images， an improving method based on the YOLOv5 network model is proposed. First， the Swin-Transformer module is incorporated into the YOLOv5 network model to model the network globally and make the full-dimensional information interactive to improve the feature extraction capability of the network； second， the loss function is optimized and the SIOU loss function is introduced to consider the vector angle problem between the real frame and the predicted frame. The comparison experiments show that the detection accuracy of the network model before and after optimization is both 95.3%. With the same detection accuracy， the recall rate of the improved network model is 91.2%， which is 0.6 percentage points higher than that before the improvement； the average detection accuracy mAP_0.5 is 95.7% after the improvement， which is 0.2 percentage points higher than that before the improvement. The results show that the improved YOLOv5 network model can improve the performance of aircraft target detection in remote sensing images to a certain extent.