A forward accuracy analysis method based on Lie-group theory and screw theory is proposed to analyze the pose error of parallel mechanisms with actuation redundancy. Both the input uncertainty and components stiffness were considered. The method was applied to study the accuracy performance of a positioning mechanism, which has the potential to be used for the positioning of satellite antenna, camera, astronomical telescope and so on. The advancement of the method is that it models the elastic deformation and passive joint motions together by motion vector and limb stiffness matrix, which greatly simplifies the constraints modeling process. Mean value and standard deviation of the pose errors under different working conditions were obtained by optimal Latin hypercube sampling method, which reduces the sampling size on the premise of ensuring the precision. Finally, the robust optimization was employed to give further insight into the effects of redundant limb and mechanism structure on the accuracy performance of the actuation-redundant mechanism.