The adsorption of organic pollutants is largely controlled by the presence of naturally occurring organic matter. The extensively used organic matter adsorption coefficients (Koc) are far from constant and tend to vary between sites. In order to identify and estimate the factors that affect this variation, a variety of soil samples from four different sites have been used to study the sorption of phenol. The soil samples were characterised by pH, electrical conductivity, loss on ignition, total organic carbon content, particle size distribution, surface area and iron oxide content. Humic,and fulvic acid along with humin were extracted from soil samples using standard techniques and analysed by 13C NMR. Standard batch procedures were used to evaluate the phenol sorption parameters in a concentration range between 1 and 500 mg/L. Acid-base titration of the soil surface was performed in order to identify possible binding sites. The study has shown that this is not possible the proportion of “glassy” to “rubbery” regions of the humic acid samples in these soil samples. The kinetic experiment showed that 97% of the phenol was absorbed after 24h. Analysis of the isotherms data showed that the Langmuir model can be used to fit data at low concentration. The wide pH variation in agricultural soils indicated different groups on the soil surface which were corroborated by the acid-base titration. The distribution and correlation of C-H, C-O, Ar, Ar-O C=O and ketone was evaluated using multivariate analysis. The results on the biplot for principal components showed that there were three trends in the variables analysed, which were C-H, C-O and Ar, indicating that aromaticity is one of the major factors driving the variability distribution.