In order to establish a daily air temperature grid data set of high precision and high spatial resolution to meet the need of public meteorological services for accurate and real-time data, the Beijing-Tianjin-Hebei Daily Air Temperature Grid Data Set (0.01°×0.01°) is obtained by spatial interpolation using ANUSPLIN based on the quality controlled daily temperature of 3 974 national and regional meteorological stations in Beijing-Tianjin-Hebei region and neighboring provinces from June to August 2018. The spatial interpolation of daily temperature data is also conducted using the IDW (inverse distance weighted), OK (ordinary kriging), and SPLINE (spline function) methods, and correlation coefficient (Corr), mean absolute error (MAE), and mean relative error (MRE) are selected as indicators to evaluate the interpolation accuracy. The results are as follows. 1) ANUSPLIN meets the requirements of spatial interpolation for accuracy and smoothness and can visually reflect the spatial distribution of temperature increasing from north to south in Beijing-Tianjin-Hebei region. 2) Among the 4 interpolation methods, the interpolation result based on ANUSPLIN is the best as the mean correlation coefficient is 0.97, the sample error within 1℃ accounts for 90.59%, MAE is 0.46 ℃, and MRE is 1.81%. 3) The regions with large interpolation errors are located in the plateau of northern Hebei Province, hilly area of Yanshan Mountains, and Taihang Mountains. High altitude and low density of stations are the main reasons for the interpolation errors. In summary, the daily air temperature grid data set using the interpolation method based on ANUSPLIN has the advantages of high resolution and low spatial interpolation error, and ANUSPLIN performs better in predicting the spatial distribution of air temperature.