This work proposes an inertial sensor-based human body structure parameters calibration methodology, aimed at being combined with a motion tracking and gesture recognition system in a virtual reality motion game context to reduce the player’s motion control learning time and improve the accuracy and ease of game operation. This proposed calibration protocol is based on the three-axis accelerometer outputs by wireless inertial sensors to calibrate user’s body parts length (including the forearm, upper arm, torso, shinbone, and whole leg) through four easy-to-perform static poses and a streamlined calibration procedure. Through this experiment, this calibration methodology proved to be a robust approach to calibrate physically normal users’ body parts length, with satisfactory calibration accuracy (overall 7.64% average calibration error rate). In addition, in order to make the calibration process more efficient, effective, and user-friendly, a calibration auxiliary system sample interface to facilitate users has been proposed in this thesis.