The thermodynamic behavior of asphalt mixtures is critical to the engineers since it directly relates to the damage in asphalt mixtures. However, most of the current research of the freeze-thaw damage of asphalt mixtures is focused on the bulk body from the macroscale and lacks a fundamental understanding of the thermodynamic behaviors of asphalt mixtures from the microscale perspective. In this paper, to identify the important thermodynamic behaviors of asphalt mixtures under freeze-thaw loading cycle, the information entropy theory, an X-ray computerized tomography (CT) scanner and digital image processing technology are employed. The voids, the average size of the voids, the connected porosity, and the void number are extracted according to the scanned images. Based on the experiments and the CT scanned images, the information entropy evolution of the asphalt mixtures under different freeze-thaw cycles is calculated and the relationship between the change of information entropy and the pore structure characteristics is established. Then, the influences of different freezing and thawing conditions on the thermodynamic behaviors of asphalt mixtures are compared. The combination of information entropy theory and CT scanning technique proposed in this paper provides an innovative approach to investigate the thermodynamics behaviors of asphalt mixtures and a new way to analyze the freeze-thaw damage in asphalt mixtures.