Abstract: To address response delays and redundant operations caused by decoupling between identification and path planning in conventional plant protection robots, an integrated algorithm framework combining disease identification with path planning, and execution under closed-loop control was proposed．First, a disease identification model optimized YOLOv8n based on GhostNet-CBAM architecture, leveraging image augmentation and heatmap generation to achieve high-precision disease detection．Next, path module employed a heatmap-guided A^* search strategy to dynamically construct high-priority task paths．Finally, a coupled control mechanism enabled coordinated identification-driven path reconfiguration and spray responses．Experiments were conducted on six categories disease classification model trained using both self-collected and publicly available datasets across, with simulation tests verifying the model's performance．Results demonstrated that coupled algorithm consistently outperforming traditional strategies across multiple metrics．Ablation tests and coupled-decoupled comparison experiments validated effectiveness and necessity of coupled algorithm．This work provided an efficient and accurate path for digital plant public protection management．