Corn,as the main grain crop,has significant significance in estimating its yield.The traditional yield estimation methods are mainly based on data sampling survey,and the accuracy of yield estimation is low.With the continuous emergence of new technologies and new methods,this paper uses UAV based on multi-spectral remote sensing technology to collect the changes of vegetation index in four key growth stages of corn,such as jointing stage,silking stage,milking stage and doughing stage,and construct an analysis model through yield,and verify and screen a reasonable model for yield estimation.It is proved that in the milk stage,the yield model based on RVI,DVI and SAVI has higher accuracy.In the wax ripening stage,the yield model based on NDVI,RVI,DVI and SAVI was higher.Through the above research,it has important guiding significance to actively combine the changes of data in the process of corn growth,facilitate efficient crop planting and estimate crop yield.