The modeling of brown rice is the basis of the further research on rice milling process. Based on the principle of the combined building of discrete fault annular section, by using high precision laser displacement sensor as detecting instrument, digital detecting device for brown rice was designed and developed to achieve a non-contact digital measurement of the discrete section of brown rice. Moving average filter and ellipse fitting were adopted according to the characteristics of the original measurement data, and the error analysis of data before and after treatment were compared and investigated. The modeling process was based on the basic theory of reverse engineering, and then three-dimensional visual digital modeling method of brown rice was discussed. It was found that the digital model based on ellipse parameters and the envelope of latitude and longitude could provide an effective reference for the study of brown rice processing.