To address the challenges in aircraft tire surface defect detection, including insufficient feature discrimination under complex textured backgrounds, high missed detection rates for small defects, and inaccurate localization of defect boundaries, this paper proposes FME-YOLO, an enhanced YOLOv8-based algorithm for common aircraft tire surface defects. First, we design the FPSC_CPCA module, which combines shared convolutions with various dilation rates and the CPCA attention mechanism to improve the extraction of fine-grained features for small targets, such as cuts and cracks. Second, we introduce the MAFPN structure, which achieves multi-scale feature fusion through cross-layer bidirectional feature transfer, improving the detection performance for small targets and defects of varying sizes. Finally, we propose the C2f_MAEFS module, which integrates multi-scale adaptive pooling, edge information enhancement, and the DSM attention mechanism to strengthen the ability to capture detailed features and improve the localization accuracy of irregularly shaped defects. Experimental results demonstrate that FME-YOLO achieves a precision of 88.0%, a recall of 75.5%, and an mAP @ 0.5 of 82.7%. These enhancements improve the performance and efficiency of aircraft tire defect detection and provide technical support for the intelligent development of aviation maintenance under the framework of smart civil aviation.