Abstract: Bird activity poses a severe threat to aviation safety, and the accurate real-time monitoring of birds around airports is a critical issue in aviation safety that needs to be resolved. To address problems such as the low location accuracy of small bird targets in infrared monitoring at airports owing to background interference and low signal-to-noise ratios, this study developed a lightweight detection model (BISTD-YOLO) based on YOLOv8n. First, to reduce the model parameters and enhance its ability to capture small target details, a lightweight multiscale convolutional module (C2f-LMSC) and an edge-guided module (C2f-E-LMSC) were designed for the backbone and neck structures, respectively. Second, the network neck and detection head structures were reconstructed to retain shallow detailed information while reducing model complexity. Subsequently, a cross-space attention-enhanced feature pyramid structure (CSAEFPN) was built. It effectively suppressed background interference and solved the feature loss problem through the synergy between the BiFPN and SimAM. Finally, the normalized Wasserstein distance loss function was introduced to replace the traditional IoU metric and optimize the location accuracy for small targets. Experiments were performed using a self-constructed AIBD. The results showed that compared to YOLOv8n, the proposed model increased the mAP@0.5 and mAP@0.5:0.95 results by 5.5% and 2.2%, respectively, decreased the number of parameters by 38.2%, and was only 4 MB in size. In generalization tests, it boosted the mAP@0.5 values by 2.5%, 1.2%, and 0.6% on NUAA-SIRST, NUDT-SIRST, and IRSTD-1k, respectively.