Oil microleakage can cause the land surface vegetation to be abnormal. An oil and gas exploration method based on the vegetation information in the hyperspectral remote sensing images is proposed. It's used to probe the effectiveness of extracting the oil and gas microleakage information, with the vegetation anomalies in the remote sensing images. A decision tree based on the vegetation index is taken to extract the anomalies areas of vegetation in the CASI images. It's shown by the experiment that there are some potential for the exploration of oil and gas in the areas covered by sparse vegetations.