Abstract: Taking the soil surface temperature and reflectance into account,NOAA/AVHRR data were fitted with the soil temperature based on different period,soil depth,land cover types,methods and regression functions.The results showed that the method built by the product of reflectance of channel 1 and brightness temperature of channel 4 of single phase NOAA/AVHRR was better than that of other methods.Its linear function was the best function fitted.It was replaceable for the models with different periods at the same soil depth.The method had passed the significant level of F0.01 and in general its relative error was less than 20% on the whole.This method also solved the disadvantages of traditional methods that couldn’t monitor soil moisture continuously and could only monitor at small scale.Furthermore,the data could be acquired easily,and the method was suited to monitor soil moisture at large scale with any level of vegetation cover.