Land surface temperature(LST) and NDVI were used for classification and soil water content(SWC) regression. Firstly, the 6 kinds of targets, dense wheat, sparse wheat, naked soil, water in ponds, silt and aquatic plants, were well classified using LST and NDVI channels. Secondly, the triangle shaped scatter plot was built and analyzed using LST and NDVI channels. Compared with the scatter plot built by red and near infrared bands, the spectral distances between different clatssifications are larger, and the samples in the same classification are more convergent. Finally, a VIT model was presented to extract SWC using LST and NDVI channel, which predicts the moisture well. The mapping of soil's moisture in the wheat area was calculated and illustrated for scientific irrigation and precise agriculture.