Abstract: Based on the fact that a launch vehicle flies in the true atmospheric density, the sounding data in a certain region from August 1, 2017 to July 31, 2019 is used to analyze the difference between the reference atmospheric density and the true atmospheric density as well as their calculated qα max of a certain launch vehicle. Finally, a revised model is established, and the results show that: 1) The altitude related variations of the atmospheric density deviations and relative deviations of atmospheric density are related to the seasons, and there are obvious changes within a year; 2) The qα max deviation shows a “M” shape change characteristic with time, with the value range of -217.1721-219.0648 Pa·rad, the average value of -9.4684 Pa·rad, and the largest percentages (20.4110%) within the range of (-20,0］; 3) The qα max deviation also has obvious characteristics of annual and seasonal variation. From January to March and November to December there are positive deviations, and from April to October there are negative deviations, i.e. there are negative deviations in spring, summer and autumn, and positive deviations in winter; 4) Using the multiple linear regression method to establish a revised model can better improve the accuracy of qα max obtained from the reference atmospheric density, with the deviation reduced from -36.1569 Pa·rad to -7.7012 Pa·rad, and the absolute difference reduced from 49.3254 Pa·rad

Key words: Launch vehicle;qα max, Reference atmospheric density, True atmospheric density, Modeling