Molecular thermodynamics studies thermodynamic problems at molecular level. The central task is to provide useful relations between molecular parameters and thermodynamic properties. To develop molecular thermodynamic models needed in engineering, we recommend an effective method by combining statistical mechanics and molecular simulation. Here the role of molecular simulation is not limited as a standard to test the reliability of the models. More directly, a few simulation results are used to fix partly the analytical form and the corresponding coefficients of the models. Two examples for establishing models for polymer systems are presented, a compressible lattice chain fluid model and an off-lattice chain fluid model. Tested by simulation results indicates that the models developed can be compared with those derived by rigorous statistical mechanics or even better. On the other hand, the models developed have simpler analytical form that can be used for engineering purposes. The advantages of the models developed are illustrated by various applications including vapor-liquid equilibria and liquid-liquid equilibria for polymer solutions, polymer blends and polyelectrolytes, density distributions of polymers confined in pores or cylinders, time-evolution of morphologies of block-copolymers, etc. Prospects for the increasing usage of molecular simulations including cell dynamics system method (CDS), dissipative particle dynamis method (DPD) and Dynamical density functional theory method (DDFT) to establish models for describing mezo structures are discussed.