Abstract: A lightweight multi-object intelligent detection model was constructed based on YOLOv8n, which was trained and validated using a self-established multi-category areca catechu plantation image dataset. This model enables effective identification of key targets such as areca catechu, weeds, operational personnel, animals, and obstacles. The results indicate that the proposed model achieves Precision and Recall of 0.842 and 0.793, respectively, with mAP@0.5 at 0.852 and mAP@0.5:0.95 at 0.496 on the custom areca catechu plantation dataset. High detection accuracy is attained for major target categories including areca catechu trees, weeds, and operational personnel. Furthermore, the model comprises only 3.01 M parameters, with an average inference time of 2.6 ms per image, demonstrating excellent real-time performance and lightweight characteristics. The methodology presented in this research effectively reduces model complexity while maintaining detection accuracy, and has been successfully deployed and validated on field equipment for understory operations.