All-sky observations of space-based microwave imager contain information of strong meteorological sensitivity related to cloud and precipitation. But the observation error in all-sky data assimilation is apparently non-Gaussian behavior which cannot meet the basic assumption that the observation error is Gaussian in the data assimilation. In this research, a “symmetric cloud” dependent all-sky observation error model developed by European Centre for Medium-Range Weather Forecasts (ECMWF) was used for Fengyun-3C microwave radiation imager (FY-3C MWRI), observations during super typhoons “Maria” and “Lekima” were selected and the all-sky observation error was analyzed after quality control and normalization. The results show that the all-sky first-guess departures normalized by the symmetric error model become to near Gaussian distribution so 〖JP2〗that the data can be used in the data assimilation. In addition, 〖JP〗all-sky approach brings 43.90%-54.63% more data than clear sky, with the biggest increases in the area of the outer spiral cloud band and cloud wall of typhoon. This result preliminarily demonstrates the assimilation potential of FY-3C MWRI all-sky observation data, and the utilization of observation data from the current Fengyun series microwave imagers〖JP2〗 is expected to improve the numerical weather prediction.