In contrast to conventional weather radar, the dual polarity weather radar transmits and receives both horizontal and vertical polarized signals, which significantly improves the accuracy of radarbased rainfall measurements and enhances the ability to classify precipitation. However, the measuring results can be affected by the radar system itself. The differential reflectivity ZDR can be estimated imprecisely, resulting in a decreased reliability of generated meteorological products. The National Weather Service made a large number of experiments on upgraded WSR88D radar and found that it is necessary to straighten the deviation made by signal differences in both transmit and receive paths. After a plenty of experiments on the Cband dual polarity vehicle Doppler radar and comparing the results with United States WSR88D test reports, it is concluded that that bias of the ZDR calibration consists mostly of 3 components which are transmit bias, receive bias, and sun measurement bias. The biases can be adjusted by using cross and parallel methods to eliminate effects of test signals, cables, and other equipment differences. Furthermore, dry aggregated snow is considered as an optimal weather target for the calibration of ZDR. The results show that the calibration method is an effective and practical way to reduce the system bias of ZDR so to ensure the precision of dual polarization radar system.