With the rapid development and application of SAR (Synthetic Aperture Radar) imaging technology, SAR-related technologies are also evolving rapidly. Among them, SAR image target detection, as an important branch, has shown significant application value. However, due to issues such as the high presence of speckle noise in SAR images, target blurring caused by target movement, and image distortion resulting from elevation differences, SAR target detection poses great difficulties, even for human visual inspection. Therefore, there is a need for an accurate and rapid detection method to complete target detection in SAR images. Since the inception of the YOLO (You Only Look Once) model in 2015, it has garnered attention for its high recognition accuracy and speed. Currently, the primarily used YOLO models include YOLOv3, YOLOv5, and more recently, YOLOv9 and YOLOv10, which were released in early 2024 and May of this year. This study utilizes YOLOv5, YOLOv9, and YOLOv10 models to detect aircraft in a SAR dataset, comparing their detection performance. The YOLOv10 model demonstrates superiority in both detection accuracy and detection rate. In my research, the YOLOv9 and YOLOv10 shows much more accuracy and frequency in detection.