A novel high-speed 3D object retrieval system is presented here. Since creating a 3D model as an input to the retrieval system is not trivial, a 2D image (photos, sketches, etc.) of the object is usually used instead. We present a new high-speed retrieval method by using skewness information. The images in the database will be represented by a skewness function. By using the proposed method, comparison between the query and all images in the database can be avoided. From the experiment, the proposed method performs 60 times faster than using only SURF method in retrieval system.