The forcing temperature in spring is the main factor that determines the flowering time of woody plants in the Northern Hemisphere. Global warming has reduced the number of chilling days in winter, which probably alters the thermal requirement of flowering. In the past 50 years, the spatio-temporal changes in the thermal requirement of spring phenology in China remain unclear. Based on the first flowering date (FFD) data of Salix babylonica and Ulmus pumila derived from China Phenological Observation Network during 1963-2018, we used three methods to calculate the thermal requirements of FFD and systematically analyzed their spatial and temporal patterns at representative sites. We also developed chilling days-thermal requirement models to quantitatively simulate the thermal requirement at each site in different years. The results showed that the thermal requirement of FFD exhibited a large spatial difference, with a relatively high value at low latitudes than that at middle latitudes. There was a significant negative exponential relationship between the average thermal requirement and chilling days across sites. The thermal requirements of FFD also changed over time. The trend in thermal requirement of Salix babylonica FFD in Guiyang, Xi'an and Mudanjiang reached 1.28-1.41 °C·d/a (P<0.01), 1.63-1.89 °C·d/a (P<0.01) and 0.12-0.58 °C·d/a (P<0.05) for the growing degree day method, respectively. The thermal requirement of Ulmus pumila FFD also increased significantly in Guiyang and Xi'an, but the trend in Mudanjiang was not significant. The decrease in the number of chilling days was the main reason for the increase in the thermal requirements. Due to the low winter temperature in Mudanjiang, the number of chilling days was large and had a small interannual variation, thus chilling days exerted no significant impact on the thermal requirement. The chilling days-thermal requirement model performed better in simulating the thermal requirement of Salix babylonica FFD, with R² of 0.54-0.66. In comparison with Salix babylonica, the model showed a relatively low precision in simulating the thermal requirement of Ulmus pumila FFD, with R² ranging from 0.33 to 0.64. Among the three methods, the thermal requirement could be better simulated by the growing degree days method compared with the growing degree days-sigmoid and growing degree hours methods. This study provides an important scientific basis for quantifying the spatio-temporal variation of the thermal requirement of flowering and for predicting the future flowering date of woody plants under the background of climate change.