Various statistical thermodynamic theories have been presented so far to understand thermodynamic properties of polymer solutions. The basic model and assumptions as well as deficiency of the theories are explained. The famous Flory-Huggins theory is favorably compared with experiment only at high polymer concentrations. The cluster expansion theory for polymer systems successfully calculates the second virial coefficient in dilute polymer solutions but not the higher virial coefficients. The renormalization-group theory makes possible the calculations of the higher virial coefficients, but it is not applicable to concentrated polymer solutions. A perturbation theory extended to polymer solutions is favorably compared with experiment except for high-molecular-weight flexible polymer solutions with low concentrations.