An integrated approach involving CE experiments, Molecular Dynamics (MD) simulations and two-dimensional NOE spectroscopy (2D-NOESY) experiments was employed to elucidate the intermolecular interactions and the separation mechanisms involved in a solvent-modified MEKC method for the simultaneous determination of diclofenac sodium and its impurities. The CE findings indicated that the addition of n-butanol (nBuOH) to the SDS micellar solution played a primary role for controlling the partitioning into the mixed micelles and the migration of the analytes and that the presence of nBuOH as cosurfactant was compulsory for achieving the complete separation of the compounds. The different capacity factors of the analytes were calculated and a change in solute association with the mixed micelle when changing the SDS/nBuOH molar ratio was highlighted. The optimal SDS/nBuOH molar ratio for the electrophoretic separation was 1:8. On the other hand, both MD simulations and NMR experiments indicated that the most favorable molar ratio for the formation of mixed SDS/nBuOH micelles was 1:2. These results suggested that probably there is an excess of nBuOH in the background electrolyte, both as free molecules and in form of aggregates, which is able to interact with the analytes, and thus may compete with mixed micelles for the considered compounds. The calculated values of gain in potential energy of the analytes when included in mixed micelles were in agreement with the observed migration order of the compounds. The role of methyl-β-cyclodextrin (MβCyD) in the background electrolyte was also investigated, since the addition of this CyD to the solvent-modified MEKC system was found to be useful to reduce the analysis time. MD simulations and 2D-NOESY spectra highlighted the formation of inclusion complexes with MβCyD not only with the analytes, but also with SDS. MβCyD may lower the availability of both SDS and nBuOH for forming micelles and mostly may compete with the mixed micelle as a second pseudostationary phase.