A brief review of modeling and simulation methods for a study of polymers at interfaces is provided. When studying truly multiscale problems as provided by realistic polymer systems, coarse graining is practically unavoidable. In this process, degrees of freedom on smaller scales are eliminated in favor of a model suitable for efficient study of the system behavior on larger length and time scales. We emphasize the need to distinguish between dynamic and static properties regarding the model validation. A model which accurately reproduces static properties may fail completely when it comes to the dynamic behavior of the system. Furthermore, we comment on the use of the Monte Carlo method in polymer science as compared to molecular dynamics simulations. Using the latter approach, we also discuss results of recent computer simulations on the properties of polymers close to solid substrates. This includes both generic features (as also observed in the case of simpler molecular models) as well as polymer specific properties. The predictive power of computer simulations is highlighted by providing experimental evidence for these observations. Some important implications of these results for an understanding of mechanical properties of thin polymer films and coatings are also worked out.