Green house gases are responsible for the global warming. Anthropic emission of Carbon dioxide represents about 70% of the total effect. One option to fight against this environmental problem is to reduce CO2 emissions in industrial effluents. Industrial processes consist in repeated cycles with dissolution of CO2 in absorbent solution and regeneration of the absorbent solution. Aqueous solutions of amines are well known to be efficient chemical solvents for the CO2 capture. A first pilot plant has already been started in Denmark (CASTOR project). This installation uses aqueous solutions of monoethanolamine to capture CO2 from power station emissions. This pilot is efficient but cannot be used as the treatment of CO2 is too expensive, and not economically usable. Reliable experimental thermodynamic data such as gas excess properties of absorbent solutions, solubility and enthalpy of dissolution of CO2 in aqueous amine solutions are needed to optimize the processes. Moreover, the development of thermodynamic models are essential to correlate and predict the vapour-liquid equilibria and enthalpic data. This presentation will focus on the acquisition of thermodynamic data and on the development of thermodynamic models needed to optimize the semi-empirical parameters used in the thermodynamic representation of the gas - absorbent systems.