Based on expounding the PV power plant state recognition, this paper extracts the relevant characteristics of the operating state of the PV station. Based on the principle of K-means clustering, through the actual operation data processing of a PV station in the city of Foshan, Guangdong Province, the corresponding feature matrix is obtained. Using K-means clustering analysis, the results show that it has important theoretical and applied value not only because the K-means clustering algorithm has a good effective on the pattern recognition of the PV station, but also can effectively solve the complexity problem of the PV station operation mode classification.