Mobile devices are too small to operate freely using their input surfaces. To solve this problem, non-contact and natural gesture interfaces have been the focus of recent research. In this paper we propose a method of estimating multi-finger position and pose for operating such devices at high speed using a single camera. Our method achieves the finger tracking based on the appearance and shape deformation model by estimating the translational movements and the degree of bent finger. The experimental results show that our method can obtain the position of the hand and the pose of the each finger within 9.7 ms.