Activated carbon fiber cloths (ACFC) have shown promising results when applied to water treatment, especially for removing organic micropollutants such as pharmaceutical compounds. Nevertheless, further investigations are required, especially considering trace concentrations, which are found in current water treatment. Until now, most studies have been carried out at relatively high concentrations (mg L−1), since the experimental and analytical methodologies are more difficult and more expensive when dealing with lower concentrations (ng L−1). Therefore, the objective of this study was to validate an extrapolation procedure from high to low concentrations, for four compounds (Carbamazepine, Diclofenac, Caffeine and Acetaminophen). For this purpose, the reliability of the usual adsorption isotherm models, when extrapolated from high (mg L−1) to low concentrations (ng L−1), was assessed as well as the influence of numerous error functions. Some isotherm models (Freundlich, Toth) and error functions (RSS, ARE) show weaknesses to be used as an adsorption isotherms at low concentrations. However, from these results, the pairing of the Langmuir-Freundlich isotherm model with Marquardt's percent standard of deviation was evidenced as the best combination model, enabling the extrapolation of adsorption capacities by orders of magnitude. Display omitted •Adsorption of pharmaceuticals onto ACFC was studied at trace concentration.•Both adsorption models and error functions were studied.•Adsorption capacities were extrapolated at trace concentrations using models.•The Langmuir-Freundlich model adjusted with a MPSD error was is the best choice.•The external mass transfer occurred outside the yarn of the ACFC.