The single chain mean field theory (SCMF) with importance sampling methodology has been applied to study homopolymer adsorption onto flat interfaces. A comparison with recent Monte Carlo data reveals that the excluded volume correlations along the chain are crucial in quantitatively reproducing the observed structure of the layer. The SCMF approach introduces a sampling of the self-avoiding one-chain conformations in a self-consistent mean-field, that allows for the local swelling of the chain. Unlike classical self-consistent field approaches, based on ideal chain statistics, the correct spatial as well as bulk density dependence of the profiles and structure are recovered by the SCMF approach.