It is widely recognised that diffusive transport of charged species in charged clay media must be consistently described for ions with different charge signs and numbers, as well as consistently with pore diffusion schemes for inert solutes that do not undergo interactions with the clay surface. This study presents a modelling approach that meets these requirements for the diffusion of tritiated water, ²²Na⁺ and ³⁶Cl⁻ tracer in heterogeneous Mesozoic sediment sequences of multiple deep-boreholes in northern Switzerland. The model uses an electrical double layer (EDL) approach and a discrete Donnan potential description of cation enrichment and concomitant anion depletion in the Donnan layer close to the basal (planar) clay surfaces to predict effective diffusion coefficients and capacity factors of the charged tracers. It incorporates diffusion data of tritiated water to determine the parameters related to the functional relationships between the accessible porosity and geometry factors with the total clay content as the main input variable. It also calculates concentration enrichment or depletion factors for mobile cationic and anionic species using relationships with the total clay content, such as the average density of surface sites or the thickness of the Donnan layer and the related volume fraction of free pore water. These factors are subsequently used to derive effective diffusion coefficients that include surface diffusion of cations and anion exclusion effects and the related capacity factors. Despite the various lithologies of the rock samples, where significant variability between carbonates and siliciclastic contents are observed, the model shows good agreement with experimental data. Discrepancies between experimental and modelled data are mainly due to inadequacies in the geometric description of diffusion pathways, rather than by potential bias in the calculation of concentration enhancement or depletion factors. However, the model approach assumes that diffusion in clay rock occurs via the pore space of clay minerals and treats all related parameters as simple average values, neglecting heterogeneity of mineral constituents and the unknown statistical distribution of the parameter values.