This paper aims at the problem of data gap when Local Analysis and Prediction System （LAPS) is merged with the raw data of new generation Doppler radar in China. Mosaic method of maximum value and distance index weights are designed for improving the original nearest neighbor algorithm in LAPS, and the reflectivity of “cone of silence” is simulated by the least square method. The results indicate that maximum value and distance index weight methods can make full use of multiple radar observed reflectivity, effectively improve the phenomenon of data gap between high elevations, and fill some grids in “cone of silence”, especially those in midtroposphere. The test of “cone of silence” reflectivity simulation by the least square method has certain effects and can simulate well the situation when observation data are enough around “cone of silence”. Thus, this research would improve the capacity of LAPS Doppler radar data assimilation, and the utilization efficiency of multiple radar observations, which would make positive effect on LAPS cloud analysis.