The aim of this paper is to identify the number of parameters of a hand model which can be computed during virtual grasping. The fingers of the user are tracked with various methods (position sensors and bending sensors) and the measured data is used to estimate the unknown parameter of the hand. An articulated hand model is constructed based on the assumption that the measured data is corrupted by noise and some of the hand's physical parameters are known. Virtual grasping and the parametric hand model are important, especially for force rendering in haptic feedback computation and to produce visually realistic results in hand-artifact interaction.