Frost and dew phenomenon observation is an important observation content of meteorological observation. The frost and dew observation uses mainly the manual method currently, which is of many drawbacks. The automatic observation principles of the capacitive frost and dew sensor are introduced, and on this basis, with focus on the test data analysis of fourmonth observation equipment from September to December 2014, the data from the device are compared with manual observation, as well as the results recognized by the standard observation device. Combining the actual situation around the installation location and the temperature and humidity information, an indepth discussion of the changing conditions of the recognition rate, failure rate, and mistake rate of reporting frost and dew and the reason that causes the recognition rate fluctuations are conducted. The experiment results show that capacitive frost and dew sensors have better recognition rate for frost and dew weather phenomena recognition in the general natural environment. On the basis of future improvement, it probably can be applied to automatic meteorological observation operationally.