This paper addresses the problem of estimating left ventricle global motion and shape from CT volumetric data based on surface modeling. We have developed an improved approach able to estimate the global motion and deformation as well as to construct a parameterized global shape of the left ventricle. The global rigid motion of the left ventricle is computed from the position and orientation change of an estimated time-varying and object-centered coordinate system while the global deformations are obtained through the fitting of global shape modeling primitives to the given digital volumes of the left ventricle. Major improvements over our initial effort include a robust algorithm for the left ventricle long axis estimation and the implementation of weighted surface fitting based on the original image statistics of the boundary points. The estimation results obtained through this improved approach match better to the apparent motion and shape pattern of the left ventricle than that of our previous effort.