Crop yield models using different VIs (vegetation index) from remote sensing data show the various precision, but each of them can provide useful information related with yield. So it is very significant how to integrate the useful information of these models. In this study, a few of typical VIs, such as NDVI (Normalized Difference Vegetation Index), SR (Simple Ratio index), TCARI/OSAVI (Trans-formed Chlorophyll Absorption Ratio Index (TCARI), and Optimized Soil-Adjusted Vegetation Index (OSAVI)), NDWI (Normalized Difference Water Index) extracted from Landsat5 TM image covering Beijing region, are used to build yield modes of winter wheat, respectively. And then the Weight Optimization Combination (WOC) method is utilized to integrate the models by calculating optimized weights to form the combining model. It is proved that the combining model with WOC exhibits better performance with the slightly higher determination coefficientR2 in comparison with each single yield models with four different VIs, respectively. The analysis of comparing the weights in the combining model with WOC indicates that the two VIs, SR and NDWI are more sensitive to winter wheat yield than the other two during the winter wheat jointing stage. The preliminary results of coupling the WOC method with remote sensing imply that WOC can be used to improve the accuracy of yield estimation based on remote sensing.