<p>This study presents numerical results on magneto-double-diffusive free convection and entropy generation analysis of alumina/water-based nanofluid within an enclosure divided by a non-Darcian porous vertical wall under internal heat generation. The porous wall, modeled using the extended Darcy–Brinkman–Forchheimer approach, operates under local thermal non-equilibrium (LTNE) conditions, with its width and position adjustable relative to the cavity’s length. The thermosolutal natural flow is induced by maintaining constant and distinct temperatures and solute concentrations on the left and right walls of the enclosure. The system is also exposed to a uniform horizontal magnetic field. The finite volume approach numerically solves the dimensionless system of controlled equations. The heat conductivity and viscosity of the nanoliquid heat transfer agent are calculated according to Corcione’s empirical correlations. Steady-state results are presented for varying Rayleigh (<i>Ra</i> = 10²–10⁵), Hartmann (<i>Ha</i> = 0–50), Lewis (<i>Le</i> = 0–10), buoyancy ratio (<i>N</i> = − 5 to 5), and heat production parameter within the porous wall (<i>Q</i> = 0–20) under LTNE porous medium conditions, analyzing hydrodynamic features, heat and solutal exchange rates, and second law behavior. Internal heat generation (<i>Q</i>) in the LTNE porous wall enhances convective motion but reduces the heat transfer rate without affecting mass transfer. It also significantly increases total entropy generation and the Bejan number, leading to higher irreversibility and lower system efficiency.</p>