The main reservoir for multicomponent industrial effluents entering the geological environment is the porous space of weakly permeable blocks of rocks.The flow of components of industrial wastes from permeable cracks or layers into a block occurs by diffusion together with the mass exchange of the liquid and solid phases due to ion exchange and the dissolution/precipitation of minerals. The most popular mass exchange model for the migration forecast is a linear sorption isotherm, where the main parameter is the distribution coefficient between the phases (K d ). In this paper we analyze the applying of this approach to complex models that take into account the multicomponent composition of the waste. The driving force of diffusion is the gradient of electrochemical potential. The first relatively simple model includes the dominant salt, and the salt of the microcomponent, whose cation can form complexes with the anion and be sorbed according to the Henry isotherm. Our second model of ion diffusion takes into account complex formation, cation exchange and interaction with minerals. The results of these models are compared with the analytical solution with a constant distribution coefficient. Liquid radioactive wastes with dangerous microcomponents, radioactive cesium and strontium, were chosen for the study. At the same time, stable isotopes of cesium and strontium exist in natural groundwater. We use the conditions of two sites of the radioactive waste disposal. The sites are chosen by the prevailing components of the composition: the waste of the first one (Siberian Chemical Works — SCW), is nitrate-sodium brine, and groundwater is fresh, the second one (Research Institute of Atomic Reactors — RIAR) has brackish waste and groundwater is chloride-sodium brine. Both models show agreement with the analytical solution (constant K d ) only for the dominant sodium nitrate. For microcomponents, cation concentrations form a maximum and anions form a minimum at the block boundary, even if the initial and boundary conditions for these components are equal. The reasons are the changing ionic strength of the solution and the change in the composition of the exchange complex. The distribution coefficient of microcomponents, calculated as the ratio of concentrations in water and rock, varies along the migration path and in time by orders of magnitude. Therefore, using of a constant Kd can lead to significant errors in the prediction of migration. However, the models of multicomponent diffusion in regional forecasts require the resource and speed of computers that far exceed the commonly used in practice studies.