In a motion tracking system, realistic presentation by a virtual avatar and detection are critical. Specifically, because the hand is an essential body part that interacts with other objects or humans in a virtual environment, accurate behavior tracking in real time is required. In this study, we propose a hand-motion detection glove using multiple tendon-inspired sensors. Remarkably, there are no sensors and electronics in the hand part of the glove. A piezoelectric sensor was positioned on the wrist part of the glove, which is connected to the tendon in the glove. The elongated piezoelectric sensors detect the angle from joint changes caused by motions of the hand. In addition, we designed a case to install multiple sensors on many joints of the fingers. To demonstrate the sensing ability for the bending motion of the finger, a series of sensor tests was conducted using a finger structure setup. We obtained the sensing results with changes in the rotational angles and frequencies. Moreover, the rotational angle can be estimated from the sensing data. Based on these results, a virtual hand system following real hand motion was constructed utilizing a motion detection glove with multiple tendon-inspired sensors. Finally, we investigated the effect of Young’s modulus, number of threads, and structure of the elastic thread to enhance the output of the sensors.