To enable efficient operation of drone detection models on resource-constrained devices while maintaining high accuracy, an improved lightweight UAV pavement damage detection algorithm, named LHP-YOLO, is proposed. This algorithm enhances YOLOv8n by focusing on two aspects: lightweight design and improved detection accuracy. First, the RG-C2f module is introduced to replace the original C2f module in YOLOv8, reducing redundant computations in the model. Additionally, a lightweight detection head, Detect-T3G, is designed to minimize the number of parameters. To further improve the model's performance in complex backgrounds and small object detection, Content-Guided Attention Fusion (CGA Fusion) is incorporated into the feature fusion structure of the YOLOv8n neck network. Moreover, Channel-Wise Knowledge Distillation (CWD) is applied to boost the accuracy of the lightweight model. Experimental results on the publicly available pavement damage dataset RDD 2022 show that, compared to the original YOLOv8n model, LHP-YOLO achieves a 3.4% increase in mAP50 while reducing the model size and computation by 41.3% and 51.8%, respectively, with parameters totaling 1.76 M and 3.9 GFLOPs. These improvements meet the real-time performance requirements for UAV-based pavement damage detection.