"The work presented in this dissertation aims at ab initio calculations of the infrared absorption cross section, heat capacity, self-diffusion coefficient, shear viscosity and thermal conductivity of a polyatomic fluid with the aid of statistical mechanics. The simulation is firstly focused on calculating single-molecule properties such as molecular structure, vibrational energy eigen values and transition dipole moments using an efficient method based on the density functional theory. Based on these ab initio results, the infrared absorption cross sections are determined by the Fermi's Golden rule, and the thermophysical properties are determined from molecular dynamics simulations. Using carbon dioxide as an example, the accuracies of the proposed calculation methods are all examined by comparing the simulated results with the experimental data in a wide range of temperature and pressure. It is found that the thermophysical properties of carbon dioxide at dilute and moderate densities can be accurately calculated without using any experimental data. It is shown that the quantum effects of molecular vibrations can be nicely accounted for by either a Monte Carlo method or an analytical correction term proposed in this work. The simulation results demonstrate the importance of considering the vibrational contribution to the thermal conductivity. It is found vibrations mainly contribute to the thermal conductivity through self-diffusion processes. It is also shown that the present method can be readily extended to calculate the temperature and density dependence of transport properties of other polyatomic fluids such as methane of which the experimental transport property data are of low accuracy or nonexistent"--Abstract, page iv.