Abstract: Unmanned aerial vehicle (UAV) thermal imaging can serve as anexcellent platform torapidly obtain the spatial distribution of crop canopy temperature (CT) for dynamic variable rate irrigation (VRI). This study aims to improve the spatial distribution accuracy of crop water deficit using UAV thermal imaging.A conversion model between SRGB and CT was established to clarify the effects of flight altitude and take-off time on the spatial distribution of CT. The experiment was conducted at the water and fertilizer integration experimental base in Dacaozhuang, Hebei Province of the North China Plain (36°28'12"N, 114°54'01"E). The UAV flight area was 1.78 ha under a three-span center-pivot irrigation system with an overhang. The experimental crops were taken as winter wheat and summer maize in 2022. Three procedures were used to determine the relationships between temperature and SRGB. Firstly, the temperature and SRGBwere displayed in the nine single photos. Secondly, the temperature and SRGB were extracted as the black body fromthe nine different temperature values, and then their photos were captured with the thermal imaging camera. Thirdly, the CT was measured with the nine stationary infrared thermometers (IRTs) that were installed in the experimental field.As such, theSRGB was obtained in these locations by the UAV thermal imaging system. In the field test, the flight date was selected in sunny and windless weather and then started at 08:00, 11: 00, 14: 00 and 17:00 with a flight altitude of 70 m and a flight period of 20 min. The flight altitude was set as 70, 90, and 110 m, according to the cruising ability of the UAV. The flight date was chosen on May 22, May 28, and June 1 at the grain-filling stage for the winter wheat,whereas,the summer maizewas on August 2, August 17, and September 13 from the twelfth leaf stage to the milk stage. The reason was that the crop canopy coverage posed some influence on the accuracy of CT during the thermal imaging system, in terms of take-off time in the experiments. The flight date was also selected on July 18, 19, and 20 at the sixth leaf stage of summer maizeover different altitudes. The results indicated that there was a significant linear relationship between SRGB and CT. The highest accuracy was achieved in the temperature of the black body, with a determination coefficient of 0.98. The take-off time of UAV thermal imaging had a great influence on the spatial distribution of CT. The spatial variation of CT showed a weak degree of variationat the take-off time of 08:00 and 17:00. At the take-off time of 11:00 and 14:00, a moderate degree of variation was found with the greatest difference in the CT and the smallest total irrigation amount in the prescription map of VRI. During the growing season of winter wheat, the total irrigation amount at 11:00 was 11.8%, 5.8%, and 20.6% lower than that at 08:00, 14:00, and 17:00, respectively. The total irrigation amountof summer maize at 14:00 was 13.5%, 2.7%, and 18.9% lower than that at 08:00, 11:00, and 17:00, respectively. The image resolution decreased,as the flight height increasedduring UAV thermal imaging. The total irrigation amount at 90 and 110 m flight height was 6.1% and 12.1% higher than that at 70 m flight height, respectively. The highest CT was achieved in the winter wheat at 11:00-15:00 and the summer maize at 11:00-16:00, indicating the optimal flying time. The findings can provide technical support to the design accuracy of variable irrigation prescription maps for the winter wheat and summer maize in North China Plain using UAV thermal imaging system.