In the field of radioactive waste management, the sorption of radionuclides measured in batch experiments has been treated predominantly in terms of empirical parameters such as distribution ratios and isotherm equations. An alternative approach is to try to understand sorption from a mechanistic view point. Such an approach has intrinsic advantages over empirical methods when trying to justify and defend sorption values chosen for safety assessment analyses and when it is necessary to predict sorption under a variety of different conditions. In this paper we propose a general approach to the modeling of sorption in natural systems based on a surface complexation mechanism. The diffuse layer model was used to describe sorption on igneous rock forming minerals. Besides the radionuclide surface complexation reactions themselves, the two main processes influencing sorption are competition and chemical speciation. These two topics are quantitatively discussed. As an example, the model and model concepts are applied to experimental results taken from the open literature on the sorption of (\mathrm{Np}\left(2 \times 10^{-11} M\right)) as a function of (\mathrm{pH}) (range 6 to 9.5 ) on a variety of minerals in an artificial deep nonsaline granite groundwater. In this modeling exercise there were no free "fitting" parameters since all the values required in the surface complexation model were taken from data, or measurements, already existing in the open literature and then fixed. No parameters were derived from the experimental data being modeled. The calculated sorption curves are compared with the measured distribution coefficients. 1993 Academic Press, Inc.