In order to test the effect of the Doppler radar data on the mesoscale model forecasting， a heavy rain event occurred on 18 July 2007 has been studied by applying the new generation Doppler raw radar data in China which contains 9 levels reflectivity and radial velocities in the ARPS (the Advanced Regional Prediction System) model. The results show that the initialization of both radar reflectivity and radial wind may significantly improve short-term quantitative precipitation forecasting by using mesoscale model. It was found that the radar radial wind adjustment may cause the small- and medium scale characteristics in the initial field and reduce the model spin-up time. As integral goes on, vapor adjustment has also an influence on the precipitation forecasting， but less significantly compared with the wind adjustment. Radar reflectivity mainly improves the humidity parameter and increases the cloud water content in the initial field. As integral begins it leads to a considerable wind adjustment. In this study， the distribution of radar echoes and water vapor centers have been predicted noticeably in the control run. The distribution of the forecasted precipitation is similar to that of the observations，but predicted radar echo intensity and correspondingly precipitation in the mesoscale model seemed to be larger.