A meaningful parametrization of an analytic equation of state can be obtained by adjusting the pure-component parameters to the vapor-pressure data and liquid-density data of the considered pure substance. For substances with no vapor-pressure data available, the parametrization of an equation of state is more subtle. In this study, we investigate two different routes to retrieve pure-component parameters for the perturbed-chain polar statistical associating fluid theory (PCP-SAFT) for such cases. In the first route, classical density-functional theory is applied, and PCP-SAFT parameters are adjusted to experimental liquid density as well as to data of surface tensions. The second approach uses results for liquid density as well as enthalpies obtained from molecular-dynamics simulations using transferable force fields. It is shown that through this route, pure-component parameters for polymeric compounds with 'vanishing' vapor pressure can be obtained without experimental data as input. We assess the approaches for several n-alkanes, 1-alkenes, and ethers as training substances. For both routes, rather satisfactory results are obtained for the correlated thermodynamic properties as well as for the predicted ones.