The potential of using zero-valent iron (ZVI) or a Fe₃O₄-loaded magnetic biochar to stabilize arsenic (As) in contaminated soil was investigated in the processes of incubation trial, chemical extraction, pot experiments with ryegrass growth. Additionally, a dry magnetic separation technique was applied to verify the possible permanent removal of As from the bulk soil. Results showed the ZVI amendment greatly reduced the As leaching, and the leached concentration became much lower than the Japanese environment standard (10 μg/L) after 180 days of incubation. Contrarily, the magnetic biochar amendment readily increased the As leachability due to the changes in pH, dissolved organic carbon, and soluble P and Si. The ZVI had a greater effect over the magnetic biochar, supported by the significantly reduced As leachability in the combined amendments. Furthermore, results from sequential extraction analysis indicate that both amendments significantly decreased the available As in (NH₄)₂SO₄ and NH₄H₂PO₄ extraction and increased the As bound to amorphous Fe oxides. But ZVI amendment alone performed better than magnetic biochar amendment alone. Plant growth experiment showed that the ZVI amendment enhanced ryegrass growth and significantly increased the ryegrass biomass. However, the magnetic biochar amendment resulted in an adverse effect on the ryegrass root growth, probably due to a marked enhancement of salinity. Meanwhile, the As uptake by ryegrass was significantly reduced in both ZVI and magnetic biochar-amended soils. Results of dry magnetic separation showed that averaged 20% and 25% of total As could be retrieved from ZVI and magnetic biochar amended soil, respectively; and the As bound to amorphous Fe oxides was the main retrieved fraction. This study indicated that ZVI or magnetic biochar could be applied as a promising amendment for reducing (phyto)availability of As in soil, and dry magnetic separation could be served as an alternative option for permanently removing As.