Finger exoskeletons and haptic devices demand an accurate, robust, and fast estimation of finger pose. We present a novel finger pose estimation method using a motion capture system. The method combines system identification and state estimation in a unified framework. The system identification stage investigates the accurate model of a finger, and the state estimation stage tracks the finger pose with Extended Kalman Filter (EKF) algorithm based on the model obtained in the system identification stage. The algorithm is validated by simulation and experiment. The experiment results show that the method can estimate the finger pose faster than 1 Khz and robustly against the measurement noise, occlusion of markers, and fast movement.