The paper first proposes a sicentific grouping method for the data of dense rainfall station network by using high density surface rainfall station network data and volume scanning data continuously gathered by radar, which were collected from the experiment of the project “the Rainstorm Flood Forecasting System in the Huang Huai River Basin”. Moreover, more than ten calibration methods which use radar and rain gauge are objectively tested and assessed with multi parameters by using grouped data. This study results show that: (1) The grouping method of rainfall station is scientific and reasonable, satisfying the request of assessment. (2) The integration method for regional rainfall estimation is better than the other 8 methods without integration. (3) Estimation accuracy by using radar features three phase distribution. (4) The range 50-100 km is the ideal for precipitation estimation while the range between 150 and 200 km is poor. (5) Accuracy and stability of region precipitation estimation method can be improved through long time practice.