Several people, particularly the elderly population, have difficulties in swallowing; therefore, to address this issue, we built an oral motion classification system based on the muscle activity pattern of the suprahyoid muscles and tested it on 12 elderly people (seven males and five females) without a history of dysphagia and five healthy young men. Surface electromyography (sEMG) signals of the suprahyoid muscles were measured using a 22-channel electrode that was designed as a thin flexible boomerang-shaped patch attached to the underside of the jaw. Six oral motions involving various tongue, jaw opening, and swallowing exercises were classified from the root mean square (RMS) features and cepstrum coefficients (CC) features of sEMG signals using a support vector machine (SVM) classifier. Results showed that the six oral motions for elderly patients were classified with an accuracy of 95.2% and 95.4% for young patients. There was no statistically significant difference in the mean classification accuracy between the two groups. We also found that the six oral motions can be classified with an accuracy of 95.2% regardless of sex, subcutaneous fat thickness on the underside of the jaw, or the level of oral function. Therefore, it appears that the system proposed here can be an effective tool for accurately measuring oral motions and it can be employed to develop an effective game-based training protocol for the elderly whose swallowing capabilities require maintenance and improvement.