The mesoscale numerical model WRF is used to simulate the No. 8 hurricane Matthew in 2016. The radar and radiometer observations are assimilated by WRF Var. With the verification to the real situation, the process of the hurricane rainstorm is well simulated by WRF in this case that it could basically show the hurricane evolution. We use the simulation results which are model outputs with high spatial and temporal resolution to do diagnostic analysis on the short term heavy rainstorm caused by Matthew, with a comparison between the best track and forecasting tracks using active and passive microwave observations from WRFDA model. In order to analyze the inner structure, the nadiral satellite-based observations were matched between the Microwave Humidity and Temperature Sounder (MWHTS) instrument aboard the FY-3C polar-orbiting platform since Sept 30, 2013 and dual-frequency radar named PR aboard GPM satellite and then separate retrievals are demonstrated in data assimilation for extreme weather with the retrieved root-mean-square errors of about 0.9 K and 17% and 10 mm/h for precipitation products, which demonstrates the impact of 118 GHz observations in data assimilation model.