Based on the L-band balloon radiosondes data in Anhui Province, the main factors such as temperature, relative humidity and wind field of ERA5 reanalysis data, the 5th generation product of the European Centre for Medium-Range Weather Forecasts (ECMWF), from 2018 to 2021 are systematically evaluated using multi-element fusion matching method, and then the applicability of ERA5 vertical elements in Anhui, China, is explored. The results show that the elements of ERA5 is consistent with those of balloon radiosondes very well, and the correlation ranging from high to low is temperature, wind and relative humidity. In the vertical direction, it seems that the contrast of each element is significantly different, with the temperature and relative humidity in the low layers being better than those in the high layers, while the wind field being the opposite. The temperature of ERA5 has the highest correlation with that of balloon radiosondes, with the overall correlation coefficient greater than 0.99. Meanwhile, there is a negative bias in temperature for the whole layer, with the absolute relative error concentrated within 30%. With the rising of the height, the negative temperature bias of ERA5 increases and the proportion of low absolute relative error decreases. The wind field of ERA5 also shows good correlation with that of balloon radiosondes, with the correlation coefficients of u and v components both greater than 0.95, the bias of -5–5 m/s, and the absolute relative error mostly within 50%. Similar to that of temperature, the correlation coefficient and the proportion of low absolute relative error of wind field increase obviously with the height, especially for the improvement of zonal wind. The relative humidity of ERA5 has a poor correlation with that of balloon radiosondes. Due to the joint influence of assimilated data and models, the correlation coefficient of relative humidity between ERA5 and radiosondes is low, the bias distribution range is wide and the concentration of absolute relative error is low, indicating the worst comparison effect. As for vertical direction, the humidity bias increases with the height, leading to the decrease of the proportion of low absolute relative error in high layers. In addition, it is shown that precipitation has a more obvious influence on relative humidity and wind field, but a weaker influence on temperature.