The spherical harmonic (SPHARM) description is a powerful surface modeling technique that can model arbitrarily shaped but simply connected three dimensional (3D) objects. Because SPHARM based 3D models can derive functional information analysis and classify different pathological symptoms, it has been used in many applications in biomedical image computing. There is an urgent requirement for efficient spatio-temporal shape modeling to represent the dynamic anatomical structures in many applications (e.g., medical image analysis, geospatial information systems). In this paper we propose a novel real spherical harmonics based spatio-temporal shape modeling method to efficiently and flexibly represent the shapes sequence of anatomical structures in medical images. Our method works well on the simply connected 3D objects and the effectiveness of our approach is demonstrated through theoretic and experimental exploration of a set of medical image applications. Furthermore, an evaluation criterion for spatio-temporal shape modeling efficiency is proposed and the comparison results showed the good performance of our method.