In the automotive wiring harness production process, the quality of connectors is critical to product performance. However, due to the current reliance on manual operations during inspection, issues such as misassembly and omissions in wiring harness connectors frequently occur. Given the limited research on automated inspection of automotive wiring harness connectors, developing a lightweight detection framework adaptable to complex industrial environments is of significant practical importance for improving inspection efficiency and accuracy. This study proposes a lightweight detection framework based on YOLO11, named SC-YOLO, specifically designed for identifying automotive wiring harness connectors. By incorporating innovative technical approaches, the framework enhances object detection performance. First, it replaces certain convolutions in the backbone network with Space-to-Depth Convolution, slightly reducing the number of parameters while improving feature extraction efficiency. Second, it introduces CBAM attention mechanism in the neck of the network, effectively suppressing redundant information and enhancing feature representation. Finally, the Inner-MPDIoU loss function is developed to address challenges in detecting small objects. Extensive object detection experiments were conducted on a dataset of connectors from actual automotive wiring harness production lines. The results demonstrate that the SC-YOLO algorithm achieves a mAP50 of 96.3%, with a precision of 86.7% and a recall rate of 97%. Compared to the original YOLO11n algorithm, these metrics improved by 0.7%, 1.6%, and 1%, respectively, confirming the advantages of SC-YOLO in improving detection quality.