The identification of line loss anomalies in low-voltage distribution networks has always been very difficult, and the large number of distributed photovoltaics connected to distribution networks has changed their power flow. This makes the identification of line loss anomalies even more difficult. In this paper, a method is proposed to identify line loss anomalies in distributed PV access station area. First, the grey correlation between PV output and line loss rate is calculated to find the correlation between PV-related factors and line loss rate for distributed PV access station area. Second, k-means clustering is carried out to select suitable indicators according to the correlation of station area line loss, and outlier detection is carried out based on the clustering results to determine whether the station area has the possibility of a line loss abnormality. Finally, by analyzing the time dispersion of the clusters where the outliers are located, it can obtain the abnormal coefficient of the station area, and judge the abnormal line loss based on that coefficient. The results show that the method can effectively identify whether the line loss of a distributed PV access station is abnormal or not through the verification analysis of typical station areas containing distributed PV.