Magnesium oxide (MgO) powder serves as the primary filler material in electric heating tubes, providing both thermal conductivity and electrical insulation. However, it faces challenges of insufficient high-temperature thermal conductivity and insulation performance. This study employs finite element simulation to systematically investigate optimization strategies for the thermal conductivity of MgO powder. Voronoi tessellation-based models were constructed for MgO powder and modified filler systems, with a focus on analyzing the effects of particle size distribution optimization, filler types, and morphology on thermal conductivity. The results demonstrate that graded powder exhibits higher packing density and thermal conductivity than single-sized powder. Under identical packing density conditions, finer particles have superior thermal conductivity to coarser particles despite the difference diminishing with increasing temperature. Thermal conductivity improves with decreasing particle size when the powder is doped with 10% spherical Si3N4 filler, and the formation of continuous thermal pathways further enhances heat transfer. For powder modified with 10% platelet h-BN filler (30 × 1 μm), the thermal conductivity is lower than unmodified MgO below 750 K due to reduced packing density. Above 750 K, the presence of a thermal skeleton improves conductivity, achieving a 34.2% increase at 1273 K compared to pure MgO.