Thermal separation processes are of great importance, because they usually cause the greatest part of costs (investment, operating) of a chemical process. For the synthesis, design and optimization of separation processes, in particular, a reliable knowledge of the real phase equilibrium behavior is necessary. If no experimental data are available, group contributions methods can be used to predict the required phase equilibria and excess properties. In recent years, Modified UNIFAC (Dortmund) has become very popular because of its reliable results obtained for different thermodynamic properties such as vapor-liquid equilibria (VLE), solid-liquid equilibria (SLE), azeotropic data, activity coefficients at infinite dilution (gamma (infinity)) and excess enthalpies (h(E)) in a wide temperature range. This paper gives an idea about the large range of applicability and the potential of Modified UNIFAC (Dortmund) for the synthesis, design and optimization of thermal separation processes and other applications of industrial interest. Furthermore, the results of an extensive model comparison for azeotropic data and excess enthalpies are presented.