Abstract: Wheat scab represents a significant threat to global wheat production, with the primary causal agents being the fungi Fusarium graminearum and Fusarium asiaticum. The dissemination of Fusarium spores represents a pivotal step in the transmission and infection of wheat scab. Consequently, the early and precise identification of pathogen spore concentrations is of paramount importance for the prompt detection of wheat scab, the safeguarding of wheat yield and quality, and the assurance of food safety. Conventional techniques, such as microscopic observation and medium isolation, are often time-consuming and require sophisticated technical expertise, posing a challenge for rapid response and the management of large-scale crop disease transmission. To address the issue of missed and erroneous detections of densely distributed small targets in scab spore images, this study has developed a small target detection model, YOLOv8-FECA. In this study, a small target detection layer with a feature map size of 160×160 was first added to the original YOLOv8 baseline model. This enhancement aimed to improve the network's ability to capture the semantic information of small targets and to enhance the accuracy of feature description. Secondly, a focus attention mechanism, FECA, was designed by combining the advantages of the CBAM and ECA modules. This module was added at the neck end of the network to reduce information loss during the downsampling process. Ultimately, in order to enhance the model's convergence capability and the precision of bounding box estimation, Wise-IoU Loss and DFL Loss were integrated as the regression loss for the bounding box. The experimental results demonstrate that Compared to the original YOLOv8 baseline model, The small target detection layer was added, the focus attention mechanism was introduced, and Wise-IoU Loss and DFL were introduced After Loss improvement, the average detection accuracy of the new YOLOv8-FECA on spore data set mAP@.5 and mAP@.5:.95 were respectively increased by 4.3% and 6.3% compared with the benchmark model Yolov8n. In terms of performance, the YOLOv8-FECA model has demonstrated an improvement in both accuracy and recall in comparison to YOLOv8n. The enhanced model demonstrates an elevated degree of accuracy and comprehensiveness in identifying wheat scab spores, with an increase of 2.6% in Precision and 3.8% in Recall. Although the frame rate per second (FPS) of YOLOv8-FECA has decreased in comparison to YOLOv8n (from 134 to 106), it nevertheless maintains a high detection speed and achieves notable enhancements in other performance indicators. This shows that YOLOv8-FECA successfully improves the detection accuracy while maintaining real-time detection capability, and mAP @ 0.5 reaches 96.8 %. The verification of the model's robustness demonstrated that accurate detection of spores is possible even in challenging scenarios, such as those involving high spore density and low light conditions. Moreover, in comparison with prevalent target detection models, including YOLOv5s, YOLOv7-tiny, SSD network, and the exemplar two-stage Faster-R-CNN, YOLOv8-FECA exhibits superior performance in terms of mAP. The results demonstrate that the improvement strategy is effective and that the new model is superior to existing methods for detecting small targets, such as scab spores. This study provides technical support for the automatic detection of wheat scab spores in the field and the early warning of scab.