Biofeedback is known to improve postural control and shorten rehabilitation periods among the young and elderly. A biofeedback system communicates with the human central nervous system through a variety of feedback modalities. Vibrotactile feedback devices are gaining attention due to their desirable characteristics and simplistic manner of presenting biofeedback. In this study, we investigate the potential of incorporating a real-time biofeedback system with artificial intelligence for wobble board training, aimed at improving ankle proprioception. The designed system utilizes vibrotactile actuators to provide forewarning for poor postural control. The biofeedback system depended on Euler angular measurements of trunk and wobble board displacements, from inertial measurement units (IMUs). A fuzzy inference system was used to determine the quality of postural control, based on IMU-acquired measurements of trunk and wobble board. The designed system integrates: 1) two IMUs, 2) a fuzzy knowledge base, and 3) a feedback-generation module. Tests were conducted in eyes-open and eyes-close conditions while standing on the wobble board to assess viability of the system in providing accurate real-time intervention. The results observed an improvement in postural control with biofeedback intervention, demonstrating successfulness of the prototype built for improving postural control in rehabilitative and preventive applications.