Based on the Doppler radar basic reflectivity and rain gauge data, the estimated and predicted precipitation products are produced by means of the precipitation algorithm, the optimal interpolation method, and the crosscorrelation method. The verification indicates that the estimated and predicted rainfall is more accurate. With the two kinds of precipitation products, the accumulated precipitation products for different periods were generated and applied in the nowcasting services of the typhoon Bilis (200604), and some other severe rainfall events (070608, 080612 and 200803) induced by lowpressure vortex shears. On the basis of the accumulated precipitation products, the yellow and orange rainstorm warnings were issued on these processes timely and effectively. The results show that the accumulated precipitation products play an important role in determining the grades and the issuing time of warning on continuous heavy precipitation processes. In the region without automatic meteorological stations, the significance is even more prominent.