The natural icing flight test is a highly difficult and high-risk subject that must be implemented for transportation aircraft certification. It needs to meet the conditions of ambient temperature, cloud Liquid Water Content (LWC) and cloud drop Median Volume Diameter (MVD) and complete the required flight manoeuvres in an environment containing a lot of supercooled liquid water. This paper uses the ground, high-altitude, satellite, radar and airborne detection data from the natural icing test flight of a domestic large transport aircraft at Yanliang Airport on 16th, 26th and 27th March 2020. The meteorological conditions of the three natural icing test flights are compared and analyzed. And the effectiveness of Li Baiping’s improved natural icing potential algorithm is verified. The results show that Li Baiping’s improved natural icing potential algorithm is more accurate for predicting icing location, but it needs further optimization in the prediction of icing intensity. Due to insufficient water vapour conditions in the west wind system, the natural icing weather conditions are worse than those in the southern branch system. Before the high-altitude cold air reaches the test flight area, the ambient temperature in the cloud is relatively high, and the icing intensity is low. The cold and warm air meets a higher LWC, a more suitable MVD, and strong icing. The research results are of reference significance for determining the window and airspace for the natural icing flight test of transportation aircraft at Yanliang Airport and the coordinated command guarantee in the future.