In order to verify the accuracy of different occultation data inversion of the wet temperature profile, it is necessary to quantify the error characteristics. The piecewise linear interpolation method, with time window 1 h and horizontal distance 100 km as the matching criterion was adopted. The average deviation and standard deviation of each interpolated height layer temperature of 0.2~30 km were statistically analyzed to explore the deviation characteristics of different occultation data. By using the complete IGRA (integrated global radiosonde archive) sounding data during 2018-2019, the deviation of different occultation data in China was evaluated. The results show that, for the above kind of occultation data, the temperature deviation is negative, and the overall average deviation is from -0.5 to 0.5 K. The absolute value of occultation data is less than 0.4 K compared with the IGRA data, and the accuracy of occultation data is equivalent to IGRA sounding station data. Through the significant detection analysis, it was found that the relative deviation of refractivity is small, combined with the statistical F=0.985 7>0.05, which shows that the detection accuracy of different masking plans has no obvious difference.