Geosorption processes play an important role as retardation mechanism in subsurface transport of organic water constituents. Geosorption behaviour of Ionisable Hydrophobic Organic compounds (IHOCs) especially influence of dissolved natural organic materials on their sorption coefficients is poorly investigated. In this study influence of pH and NOM on the sorption of three different phenols (2-methyl-4,6-dinitrophenol, 2,4,6-trichlorophenol and pentachlorophenol) on a natural sandy aquifer material collected from a bank filtration site was studied in flow through column experiments. In first step the sorption coefficients of investigated phenols from NOM-free solutions with different pH values were determined. Single as well as multi (mixed) solutes experiments confirmed that there was no competition effect among the phenols investigated. An increasing sorption/retardation was observed with decreasing pH. A linear model was used to resolve the apparent sorption coefficient into an ionised species and neutral species component. This model allows a prediction of apparent sorption coefficients for any pH value. Analysis of data also revealed that the anion of 2-M-4,6-DNP show no significant sorption and a simple model based on the neutral form sorption coefficient was useful in predicting the apparent sorption coefficient over the entire pH range investigated. For 2,4,6-TCP and PCP, the anion sorption is much lower than that of the neutral species but cannot be neglected completely. For PCP, the anion sorption is even as high as the sorption of neutral 2,4,6-TCP. Thus for such hydrophobic phenols prediction of sorption based only on the partitioning of non-dissociated species may lead to an error, especially at high pH values (> pKa), where ionised form is present predominantly. On the other hand prediction of sorption based on the partitioning of non-dissociated species is useful at pH < pKa or in cases where the anion sorption can be neglected. The advantage of this prediction is that it allows predicting the pH-dependent sorption only on the basis of LFER correlations. It was shown that the organic carbon normalized sorption coefficients of the neutral forms of the phenols can be predicted better by solubility based regression correlations than by n-octanol-water partition coefficient based correlations. In second step the apparent sorption coefficients of investigated phenols from NOM-containing solutions (mixed solution of phenols dissolved in natural river water) with different pH values were determined; the sorption coefficients obtained from this step were compared with the results of the first step examinations. Analysis of data revealed that binding of 2-M-4,6-DNP by dissolved humic substances (DHS) reduces the sorption coefficient significantly. In contrast to it, the sorption of the more hydrophobic chlorinated phenols was not affected by DHS. A strong reduction of sorption onto aquifer material caused by complex formation with DHS was also observed for a second nitrophenol (2,4-DNP) and seems to be typically for nitrophenols. A combined sorption and complex formation model was proposed which could be used successfully to describe the effect of pH and NOM concentration on sorption reduction and to estimate binding constants. The binding constant found for 2-M-4,6-DNP is much higher than those reported by Ohlenbusch and Frimmel, 2001, for chlorophenols associated with Aldrich humic acid. This can be interpreted as a result of specific interactions between the nitrophenol and DHS. This phenomenon may cause an increase of the amount of nitrophenols, which can be transported with the groundwater flow and has to be considered in transport modelling, especially in cases where pH is lower than pKa. The models developed here are a useful tool to describe the influence of pH and NOM on geosorption processes and to estimate the Kd values which have to be used in transport models.