At present, there are five S-band polarimetric weather radars operating in Guangdong province. Due to the technical limitation of the current polarimetric weather radars, the raw polarimetric variables are still unstable, andthe data availability and the main influencing factors must be analyzed in order to ensure the reliability of radar products. Based on the continuous observation data of the dual polarization radar in Guangzhoufrom July to August in 2016, the effects of signal-to-noise ratios (SNR) on cross-correlation coefficients at zero lag (rho_(HV) (0) ) and differential reflectivity (Z_(DR)) were analyzed.The result shows that only when SNR reaches 20.0 dB or above, the radar data is credible, and the SNR threshold value of the data can be decreased to 17.0 dB after noise correction. The relation between reflectivity (Z_H) and Z_(DR) after noise correction was also analyzed and compared with the fitting results of the raindrops spectrum and the simulated data. The result shows that the measured value of Z_(DR) of a polarimetricradar in Guangzhou is less than both of the fitting results and the simulated data. After the micro-raindrop method index was adjustedaccording to the climatic characteristics of Guangdong, the system bias of Z_(DR), initial differential phase(Phi_(DP)(0)) with the different azimuth and time were further analyzed. It is found that Z_(DR) fluctuates between -0.29 dB and -0.22 dB with the azimuth, the average deviation is -0.09 dB after excluding theblockage; and Phi_(DP)(0) fluctuates about 4° with the azimuth, and the system biases of Z_(DR), Phi_(DP)(0) fluctuatewith time obviously. In the end, the Z_(DR) was corrected in terms of noise and system biases, and the result shows the quality of Z_(DR) has been improved. These preliminary results are useful for polarization radar data applications.