We studied the effect of different doses (0.5%, 2% and 5% (w/w)) of ammonium thiosulfate on mercury (Hg) speciation fractionation following its addition to the soil, as well as its accumulation by oilseed rape (Brassica napus L.), corn (Zea mays L.), and sweet potato (Ipomoea batatas L.), and compared them to a non-treated control in a historically polluted soil. The oilseed rape, corn, and sweet potato were planted consecutively in the same soils on days 30, 191, and 276, respectively after the addition of thiosulfate to the soil. The key results showed that bioavailable Hg contents in the rhizosphere soils ranged from 0.18 to 2.54 μg kg⁻¹, 0.28 to 2.77 μg kg⁻¹, and 0.24 to 2.22 μg kg⁻¹, respectively, for the 0.5%, 2% and 5% thiosulfate treatments, which were close to the control soil (0.25 to 1.98 μg kg⁻¹). The Hg L₃-edge X-ray absorption near edge structure (XANES) results showed a tendency of the Hg speciation to transform from the Hg(SR)₂ (initial soil, 56%; day-191 soil, 43%; day-276 soil, 46%, and day-356 soil, 16%) to nano particulated HgS (initial soil, 26%; day-191 soil, 42%; day-276 soil, 42%, and day-356 soil, 73%) with time in the soil treated with a 5% dose of thiosulfate. The Hg contents in the tissues of the crops, except for oilseed rape, were slightly affected by the addition of thiosulfate to the soil at all dosages, compared to the control. The addition of thiosulfate did not induce the movement of bioavailable Hg to the lower layer of the soil profile. We conclude a promotion of Hg immobilization by thiosulfate in the soil for over one year, offering a promising method for in-situ Hg remediation at Hg mining regions in China.