Wire endpoint detection is crucial for experimental wiring inspection using artificial intelligence. To address the problem that the existing wire endpoint detection model has large parameters and is difficult to deploy on mobile terminals, this paper proposes a lightweight wire endpoint detection method based on YOLOv5. CSPDarkNet was replaced by PPLCNet as the backbone network to maintain high detection accuracy while reducing the complexity of the model. ConvNeXt Block was integrated into the feature fusion part to accelerate network extraction and fusion of complex target feature information, and enhance the feature extraction capability of the target. The convolutional layer in the neck network is replaced by a lighter convolutional GSConv to reduce the computational cost. The experimental results show that the calculation amount, parameter number and model volume of the improved model are reduced by 66.6%, 68.4% and 65.2% compared with the unimproved model, and the mAP is increased by 0.9%, which ensures the lightweight and improves the detection accuracy.