Current technologies for grading betel nuts remain rudimentary and heavily rely on manual inspection, which results in no assurance over grading quality. This paper describes a betel nut grading method based on the computer vision technology, which comprises image capture and pre-processing to collect characteristics such as color, shape, and texture. It is observed that combining both color and shape characteristics leads to satisfactory grading results. This method can reach a recognition rate over 90.38% following principle components analysis (PCA) and a support vector machine (SVM) algorithm.