Apple leaf diseases significantly impact fruit quality and yield. To overcome the limitations of existing detection methods in accuracy, real-time performance, and robustness under complex noise conditions, this study proposes SMPD-YOLO, a lightweight and real-time apple leaf disease detection model. Specifically, a Spatial Pyramid Pooling Fast Cross Stage Partial CSP (SPPFCSPC) module is integrated into the backbone network to enhance feature fusion. The Minimum Point Distance-IoU (MPD-IoU) is employed as the bounding box regression loss to improve model precision and accelerate convergence. The model is further compressed and computational complexity reduced via Layer Adaptive Magnitude-based Pruning (LAMP), while channel-wise knowledge distillation (CWD) is applied to boost detection performance. Experimental results demonstrate that SMPD-YOLO achieves a mean Average Precision (mAP@0.5) of 90.20% and a frame rate of 133.3 frames per second (FPS), with a model size of 5.0 MB and 7.3 GFLOPs. Additionally, SMPD-YOLO maintains strong robustness under challenging conditions, including strong illumination, low light, and image blur. Overall, the model combines high accuracy, lightweight design, and real-time performance, providing a lightweight design approach for achieving efficient leaf disease detection on resource-constrained equipment.