Complex mixtures of contaminants with potential adverse effects on human health and wildlife are found in the environment and in the food chain. These mixtures include numerous anthropogenic compounds of various origins and structures, which may behave as endocrine disruptors. Mixture's complexity is further enhanced by biotic and abiotic transformations. It is therefore necessary to develop new strategies allowing the identification of the structure of known, as well as unknown, nuclear receptor (NR) ligands present in complex matrices. We explored the possibility to use NR-based affinity columns to characterize the presence of bioactive molecules in environmental complex mixtures. Estrogen receptor α (ERα)-based affinity columns were used to trap and purify estrogenic substances present in surface sediment samples collected in a French river under mixed anthropogenic pressure. We combined biological, biochemical and analytical approaches to characterize the structure of ligands retained on columns and demonstrate the presence of known active molecules such as bisphenol A and octylphenol, but also of unexpected ERα ligands (n-butylparaben, hydroxyl-methyl-benzofuranone). High resolution mass spectrometry results demonstrate that ERα affinity columns can be used for the isolation, purification and identification of known as well as unknown estrogenic contaminants present in complex matrices.