The line loss analysis of the low-voltage distribution network station area is of great significance to discover and solve the problem of abnormal line loss, reduce electricity loss and fine management of users. Based on the real data collected by the full-event electricity information acquisition system, this paper proposes a novel line loss diagnosis method for low voltage station area. In this method, the grid diagnostic rules are used to analyze the quality of the collected raw data. By extracting and classifying of the line loss characteristics of the station area, a binary K-means clustering diagnostic algorithm is established based on voltage information, as well as a global search diagnostic algorithm based on power information, which realizes the rapid positioning of abnormal meters and the treatment of line loss anomalies in the station area. The results show that the proposed method has high accuracy and certain practicability by eliminating abnormal meters and actual tests.