Sodium volume modeling during hemodialysis encounters several difficulties. First, the actual sodium distribution volume is the extracellular water, whereas the ultrafiltration flow reflects the variation of total body water. Thus, a two-pool model must be considered. This will complicate the model by increasing the number of parameters and boundary conditions. An alternative is to consider the total body water as the apparent distribution volume of loaded or removed sodium, which leads to a single-pool model. Second, convective sodium transfer induced by ultrafiltration is not negligible compared with diffusive sodium transfer. Therefore, sodium transfer modeling must simultaneously take into account the diffusive and the convective part, with the coupling part related to both processes. Third, the Donnan effect due to nondiffusible anionic plasma proteins modifies the sodium transfer through the membrane. Adequate sodium volume modeling should be a compromise between oversimplification, resulting in discrepancies between calculated values and experimental data, and overcomplexity, involving a great number of parameters and boundary conditions, which leads to a model unsuitable for clinical application. A single-pool model is proposed with only one parameter that is estimated during the first period of the hemodialysis session.