In modern agriculture, the monitoring and detection of rice pests is crucial for ensuring food security. However, traditional manual detection methods are time-consuming and difficult to scale. In response, this paper proposes an improved YOLOv8 model for accurately identifying and detecting pests in rice crops. By incorporating attention mechanisms and the BiFPN feature fusion module into the model, the ability to recognize target objects and capture local features has been significantly enhanced. Experimental results show that the proposed model outperforms traditional YOLO models in terms of detection accuracy, speed, and recall rate, demonstrating its high practical value.