Along with steam injection, CO2 flooding has been proved to be among the most promising EOR methods. In CO2 flooding CO2 acts as a solvent, picks up lighter hydrocarbon components, resulting in swelling of oil. These effects reduce the oil viscosity and increase oil saturation. Thereby, the oil mobility improves, and a highly efficient oil displacement process can be achieved. In all CO2 recovery processes, the long term dissolution behavior of CO2 in the reservoir fluid is diffusion controlled. In processes where the diffusion rates are low, non-equilibrium effects may reduce the displacement efficiency. Therefore, the diffusion coefficient is very important for the performance prediction. The diffusion rate of CO2 in reservoir oil and brine is required in reservoir simulation to predict the oil recovery for CO2 miscible flooding. To our knowledge, only few studies have been carried out to experimentally investigate the gas diffusion coefficient of CO2 gas into liquid phases. This thesis describes experiments done in a cylindrical PVT-cell, where a certain volume of CO2 gas was brought into contact with brine or oil. The equilibrium diffusion model by Civan et. al. is applied for determination of the gas-diffusion of CO2 from pressure decline by dissolution of CO2 in liquids (oil or brine water). By knowing the rate of dissolution of CO2, the amount of CO2 (including the loss of CO2 to the coexisting aqueous phase) for recovery process can be predicted more accurately.