By using the L-band radiosonde data of Shanghai Baoshan Station during the winter from 2019 to 2021, the FY-4A cloud top temperature (CTT) product is evaluated. The results indicate that the CTT of the single-layer cloud can be reflected well by the FY-4A CTT product, while the CTT of multi-layer clouds is generally overestimated with the mean bias larger than 14℃. And the FY-4A CTT product has a small mean bias of approximately 3℃ compared to the sounding observation when the observed cloud top height is below 6 km or the observed cloud top temperature is higher than -20℃. Moreover, based on the surface observation in the Yangtze River Delta Region in the winter of 2021 and 2022, the relationship between CTT and precipitation types in winter seasons is statistically analyzed and further verified by typical cases. The results reveal that a low CTT is one of the main necessary conditions for snowfall, as most snowfall events occur when the CTT is below -12℃. As the FY-4A CTT product has a certain accuracy with a high spatial and temporal resolution, this product could have great application potential in iden-tifying precipitation types.