According to prevention level of grain storage requirements and granary actual situation, a novel grain storage insect image for automatic identification algorithm was presented based on features point and stereo matching. Firstly, pyramid delaminating was used to preprocess target image. Then, image normalized could be obtained according to formula. Secondly, the SIFT features points could be extracted from reference image and target image. Finally, according to epipolar constraint and the straight line slope of pass through features matching points, The paper could be successful to automatic identify grain storage insect. The results of experiment with the really grain storage insect images shown the proposed algorithm was effective and could obtain high recognition rate.