Protein-surface interactions are of great importance in a wide variety of applications in biomedical engineering and biotechnology, including medical implants, biocatalysis, immobilized-enzyme bioreactors, biosensors, bioseparations, and bioanalytical systems. While this is well recognized, very little is understood regarding how to design surfaces to optimally control protein adsorption behavior. To address this limitation, we are working on the development of molecular simulation methods to accurately predict protein-surface interactions at the atomic level. We have found that this type of molecular system is sufficiently unique that molecular simulation methods cannot simply be borrowed from other applications; but rather, they must be critically evaluated and often modified to accurately represent adsorption behavior. In this talk, I will address four major areas that we have identified as being particularly important for the simulation of protein-surface interactions, and I will present our approaches to address each of these areas. These are (1) the general methods that are needed to properly simulate protein-surface interactions, (2) the suitability of a force field to represent protein-surface interactions, (3) the adequate treatment of solvation effects, and (4) the need for advanced sampling methods for large molecular systems. I will present an overview of our efforts to address each of these key areas. We are developing a hybrid force field program that enables multiple force fields to be used in a single simulation to represent different phases of a system (e.g., solid surface, solution, and the interphase between them), methods to enable pressure to be properly monitored and controlled in a simulation with constrained atoms, and how electrostatic effects should be represented for surfaces with high charge density when using periodic boundary conditions. We have generated a large experimental benchmark data set for peptide-surface interactions for use for force field evaluation, modification, and validation purposes along with simulation methods to calculate adsorption free energy for comparison with this data set. Regarding solvation effects, we have found that existing implicit solvation methods are completely unsatisfactory at this time and must be redeveloped before use in protein adsorption simulations. Finally, we are also developing advanced sampling methods for large molecular systems to efficiently overcome energy barriers that often cause simulations to become trapped in local low-energy states and prevent proper exploration of the relevant phase space of the molecular system.