Based on the basic method of consistency evaluation between 〖JP2〗radar base data, this article analyzes〖JP〗 the influence of radar altitude difference, terrain blockage, noise, congestion, temporal and spatial overlap rate, attenuation, etc. on consistency evaluation algorithm result quality, and takes corresponding improvement measures, that is, using radar beam blockage percentage data to correct or directly eliminate overlapping points with blockage, eliminating the overlapping points with low SNR, incomplete filling and abnormality, adding the factors of temporal spatial overlap rate, and analyzing the impact of different threshold temporal and spatial overlap rates on the evaluation results. The improvement measures have improved the quality of the deviation mean and standard deviation in the consistency evaluation results. The improved algorithm is used to evaluate the radar system deviation of Yueyang Station and the consistency of radar echoes in Jiangsu and Anhui regions. The results show that the deviation value between the Yueyang Radar and the surrounding radars has obviously decreased after being maintained, returning to the normal level, and the standard deviation value has also decreased slightly. The average value of deviations between radars in Jiangsu and Anhui is mostly within 2 dB. The vector triangle composed of the deviation mean has a good closure, with standard deviation between 2.9 dB and 3.3 dB and correlation coefficient between 0.40 and 0.55.