In order to verify the performance of the raindropspectrum precipitation phenomenon sensor, the PARSIVEL raindropspectrum precipitation phenomenon sensor deployed at Jiangning by the Nanjing Meteorological Service is used to collate the observation data of weather phenomena from January to November 2018, and the performance of the sensor for different types of precipitation is analyzed. The conclusions are as follows: (1) The raindropspectrum precipitation phenomenon sensor has good consistency with manual observations in general, but there are some differences in precipitation duration and precipitation types. (2) In the daytime, the total precipitation duration of manual observation is obviously higher than that of the sensor, and the main reason is that manual observation might ignore the time between precipitation events, while the sensor might miss very light precipitation; the precipitation classification of manual observation is rough, and the fine changes of precipitation may be neglected, while the sensor is able to detect the shortterm changes in rainfall, the conversion of drizzle and rainfall, the conversion of rainfall and snow, etc. (3) At night, the manual observation may ignore a certain drizzle or rainfall, whereas the sensor tends to miss the light rainfall, and rains might be misjudged as graupels or snowfall. Therefore, the algorithm for classifying precipitation still needs to be improved.