Abstract: Five kinds of remote sensing models were established for estimating the forage yield of the meadow in Northern China based on NOAA/AVHRR NDVI data, and the forage yield from the ground observation. The correlation coefficients of the five models were more than 0.7, which showed that the models were effective. Because the data of the forage yield were measured in the provinces such as Inner Mongolia, Gansu, Ningxia, Xinjiang, Qinghai and Tibet where the meadow was the main vegetation type, they were representative for the whole grassland in Northern China. So the models could extend the study area accordingly. The spatial analysis of the model results showed that the fertile meadows in Northern China lied in the eastern Inner Mongolia, Qilian Mountain in Guansu, Tianshan Mountain and Altai Mountain in Xinjiang, and the eastern Qinghai-Tibet Plateau. In this sense, the models could be applied in the survey of the resource of grassland. The time analysis of the model results revealed that the forage yield was mutative with the climate. So, the models could be applied to estimate the loss because of the climate disaster.