Abstract: Aiming at the problems of high false detection rate of pedestrians in a complex background,a low detection accuracy of dense pedestrians and missed detection of pedestrians with small distant targets in the infrared image pedestrian detection algorithm,this paper proposes an infrared pedestrian detection algorithm based on attention and feature fusion module-you only look once (AFFM-YOLO).Firstly,an attention feature extraction module (AFEM) is proposed,which is integrated into the backbone of the network to suppress irrelevant background information and enhance the extraction of key feature information.Secondly,a multi-scale feature fusion module (MFFM) is designed,which is embedded in the neck of the network to realize the effective fusion of feature information at different scales.Then,a large-scale detection layer is added to strengthen the feature extraction ability of the target detector for pedestrians with small targets in a long distance.Finally,the experimental results on FLIR data set show that the average accuracy of AFFM-YOLO reaches 89.1,which is 2.4% higher than that of the baseline algorithm YOLOv5.AFFM-YOLO has a better performance,with a significant improvement of the pedestrian detection effect in infrared images.