This study is base on machine vision technology，choose the soybean which from Heilongjiang，America，Brazil and Argentina，through the MATLAB image processing，combined with soybean moisture content，in order to be able to accurately predict the moisture content in soybean.Finally，through the MATLAB GUI（graphical user interface）design and develop a set of online moisture content of soybean moisture measurement software.Through the comparison of predictive value and the true value of the soybean，the prediction accuracy of this software could run up to 97% under the same batches.This software provide a reliable theoretical basis for the online determination of moisture of grain.