To address the issue of misjudgment in meter wiring errors caused by light loads and reactive power overcompensation un-der high-proportion distributed energy integration, an intelligent identification method based on XGBoost is proposed. The inte-gration of distributed energy resources significantly alters grid load characteristics, especially under conditions of fluctuating renewable energy sources such as photovoltaics and wind pow-er, leading to light load scenarios. This method analyzes features such as current and power factor in light load and reactive power overcompensation scenar-ios, combining actual collected data with generated data to build a classification model. The model effectively distin-guishes normal operating conditions from wir-ing errors. Validation results demonstrate that the model achieves over 98% accuracy in detecting wiring errors, as measured by metrics such as accuracy, precision, recall, and F1-score, sig-nificantly reducing misjudgment rates. The findings show that this approach can effectively adapt to the complex power load scenarios caused by high-proportion distributed energy integration, providing an efficient and reliable solu-tion for intel-ligent identification of wiring errors in power systems.