Abstract: To address the low troubleshooting efficiency, poor localization accuracy, and high cost in line loss management of low-voltage distribution networks in power supply stations, this paper proposes a lean line loss management technology based on virtual segmentation and anomaly attribution. Based on existing collected data, including transformer area master meter data, smart meter data, and time-series voltage data, this technology constructs a process including anomaly screening, data preprocessing, virtual segmentation, segment-level scoring, anomaly attribution, and closed-loop management. It realizes virtual segmentation of users in transformer areas by integrating voltage similarity with archive information, establishes a five-dimensional segment-level anomaly scoring model to identify high-risk sections, and completes five types of attribution based on a rule base, including data acquisition, archive records, metering, technical high line loss, and abnormal electricity consumption. Experimental results show that the anomaly identification accuracy of this method reaches 92.3%, and the anomaly localization range compression rate reaches 65.3%. The proposed method significantly narrows the scope of on-site verification, reduces management costs, and provides a practical solution for lightweight and precise line loss management in low-voltage transformer areas.