High-frequency radiometric calibration can track and correct the on-orbit state of the remote sensors in time to ensure the accuracy of observation data. In order to improve the accuracy and data consistency between the different remote sensors, the radiometric benchmark satellites have been developed in recent years. However, during the radiometric transmission of the radiometric benchmark satellite, due to the different orbits between the benchmark satellite and the remote sensor to be calibrated, the observation angles between the satellites are quite different, so the calibration frequency is limited under the condition of Simultaneously Nadir Observation (SNO) cross calibration. In order to improve the calibration frequency and reduce the impact of angle matching on calibration accuracy, a cross calibration method based on the multi-angle apparent reflectance model is proposed in this paper. The multi-angle apparent reflectance model of the stable target site is constructed to carry out the cross calibration of the benchmark satellite. The stable site used for the cross calibration is selected, and the multi-angle model of the site is constructed by grouping the observation angles, solar angles, and apparent reflectance of Terra/MODIS and AQUA/MODIS for 11 years from 2008 to 2019 under the multi-angle and high-precision remote sensors. The multi-angle model is suitable for the large angle cross calibration in the range of 0°~60° zenith angle. Based on the pseudo-invariance property of the stable site, the spectral matching function suitable for 400~2 400 nm band is constructed by using the average spectral apparent reflectance data of Hyperion, ignoring the variation of the time-dimensional. In the process of the benchmark satellite cross calibration, regional matching, time matching, and cloud removal are carried out for the reference sensor and the sensor to be calibrated at first. Then, the apparent reflectance measured by the reference sensor and the sensor to be calibrated is obtained using the model to complete angle matching and spectral matching. Finally, the calibration coefficients are calculated by combining the DN value of the sensor to be calibrated. According to the calibration principle in this paper, a multi-angle apparent reflectance model of the pseudo-invariance field Libya 4, which is recommended by CEOS, was constructed. Using AQUA/MODIS as the benchmark sensor, 205 times cross-calibration applications were performed on MERSI II of the Fengyun-3D satellite at six solar reflection bands from 2019 to 2020. The calibration coefficient and in-orbit variation trend of FY3D/MERSI II high frequency were obtained. In order to verify the accuracy of the cross calibration method in this study, 26 times a comparison of results was compared with SNO cross calibration method with strict angle limitation. The two methods have good consistency in the measurement trend. Compared with SNO cross calibration method, the calibration frequency increased from 26 times to 205 times, and the average relative deviation was less than 2.1%, which proved that the multi-angle hyperspectral apparent reflectance model constructed in this study could be better applied to cross calibration under the condition of large angle difference. The apparent reflectance model of the stable site is constructed by using the long-sequence, multi-angle and hyperspectral high-precision satellite data. The apparent reflectance model is suitable for the multi-angle cross calibration of the remote sensors within 60° observation angle and observation band within 400~2 400 nm. The calibration method in this study can solve the dependence on the multi-angle ground site data from the manual measurement, significantly improve the calibration frequency of the multi-type satellites. The calibration method timely track and correct the variation trend of load in orbit, and improve the quality of remote sensing data. However, the accuracy of the single point calibration and the seasonal fluctuations need to be further analyzed and improved. Through the application and comparison verification of more stable sites and more remote sensors, this method will promote the application of the multi-satellite quantitative remote sensing.