Abstract: To achieve more accurate detection of tea diseases， a feature map was reexamined from three distinct perspectives： channel‑width， channel‑height and width‑height. Based on those perspectives， a novel residual multi-scale enhanced attention module was proposed， and the semantic information and spatial characteristics of disease images were efficiently extracted. Meanwhile， a new triple attention-based feature fusion enhancement module was designed by integrating the above three perspectives. Channel and spatial features were deeply fused， which significantly improved the model's ability to distinguish between leaf backgrounds and diseases. Experimental results showed that the improved model achieved excellent performance， with a mean average precision （mAP） of 95.2%， a precision of 93.1%， and a recall of 92.0%， which was superior to existing detection methods.