Research shows that regular physical exercise is essential to improving the overall health of older adults. However, many older adults find it difficult to stick to a dedicated exercise program. To address this problem, this paper selects reclining chairs that are common in daily life in order to integrate exercise into daily life. Combined with an intelligent system, the posture of the person is inferred and the intensity of the exercise is adjusted according to the different strength and physical state of each elderly user. This paper generates a digital twin system by using pressure sensors and IMU (inertial measurement unit) sensors embedded in the reclining chair to detect the posture of the elderly in the reclining chair. Compared with traditional visual methods, visual methods have the risk of privacy issues, and the amount of force exerted by the elderly on the reclining chair is difficult to judge visually. By adjusting the difficulty of the exercise according to the user’s current physical strength, this system can help the elderly maintain their fitness habits and improve their quality of life.