Latent heat storage is widely investigated by the researchers due to its high volumetric energy density which makes it possible to largely reduce the energy storage cost. However, the phase change materials are known to suffer from poor thermal conductivity which greatly limits its use in the industry. A large amount of work has been carried out to enhance the heat transfer capability of the latent heat storage system. From a structural perspective, embedding of fin structure into the heat storage tank is considered as an effective way to lead to the overall heat transfer enhancement. For the moment, shape optimization and sizing optimization are the two most common optimization methods that are used to find the efficient finned structure in a heat storage tank. However, by predefining the shape of the geometry, the shape optimization and sizing optimization have more constraints when performing the optimization which limits the possibilities of a design change. Different from these two optimization approaches mentioned above, the topology optimization requires only few constraints when performing the optimization which enables dramatic design change without predefining the shape of the geometry. The core problem solved by the topology optimization is about the distribution of the materials and their topological connection within the design field. By topology optimization, high thermal conductive materials can be distributed in the heat storage tank in a way which maximizes the overall heat transfer capability of the heat storage system. The topology optimization of a classic tube-and-shell latent heat storage tank is studied in this paper to enhance the overall heat transfer capability. By combining the topology optimization theory and the classic finite element method, a 2-D heat storage tank model has been built for the optimization of the fin structure. Besides, a comparison between the topology optimized fin structure and other typical types of fin structures is carried out. The numerical simulation is also performed considering the effect of natural convection to see its impact on the design change of final result. Furthermore, the current research on topology optimization focuses mainly on the design phase. Few studies have been done to validate numerically the reliability of the reconstructed topology optimized design which is necessary before an experimental validation. Hence, in this paper, a numerical reconstruction of the fin structures is carried out and the validation of the result is performed. Several results could be drawn from the current research: The topology optimization shows its advantage over common fin structure design; The influence of natural convection on optimization has been investigated and analyzed; The result has been reconstructed in common CAD form and corresponding validation has been performed which serves for the preparation of the upcoming experimental investigation.