Aiming at the problems that the appearance characteristics of road cracks are easily disturbed by the environment, the missed detection rate of small cracks is high, and the computing resources of detection equipment are limited, a lightweight detection model MCA-YOLO-A is proposed. Based on YOLOv8n, this model replaces the backbone network with a lighter MobileNetV3 feature extraction network, and integrates the CA attention module with accurate spatial information to improve the feature extraction ability. At the same time, the Alpha-IOU loss function suitable for lightweight network is introduced, which improves the overall performance of the network. In addition, a small target detection layer is added to improve the identification accuracy of small cracks. The mAP_0.5 and F1 Score of MCA-YOLO-A model on road crack data sets is 0.930 and 0.893, respectively, which is 7.0% and 9.7% higher than that of the original YOLOv8n model, and the parameter quantity is only 6.0M, which is 4.8% lower, and the detection speed reaches 95 frames/s.. The experimental results show that the model has high accuracy, small model and strong generalization ability, and is more suitable for deployment in scenes with limited computing resources such as embedded systems and mobile devices.