Petroleum hydrocarbon, a complex mixture including aliphatic and aromatic hydrocarbons, is known to have negative effects on the environment. We determined the effectiveness of persulfate (PS) (5% w/w) activated using 1% (w/w) of various types of soil minerals, goethite (α-FeOOH), hematite (Fe2O3), magnetite (Fe3O4), maghemite (γ- Fe2O3), manganese oxide (MnO2), and zeolite (clinoptilolite) to treat petroleum hydrocarbon-contaminated soil. Total petroleum hydrocarbon (TPH) was 4200 ± 124 mg kg−1. The TPH removal efficiency was in the order: Fe3O4 > MnO2 > γ-Fe2O3 > Fe2O3 > α-FeOOH > clinoptilolite. When the PS dosage and the moisture content of the soil increased, the degradation rate (k obs) of TPH removal increased; on the other hand, lowering the pH increased the k obs of TPH removal. The PS oxidation of TPH was optimized using response surface methodology (RSM). The interactive effects of three factors, namely, persulfate dosage (X 1), pH (X 2), and soil moisture content (X 3), were investigated. The optimum removal of TPH by PS oxidation activated using 1% (w/w) magnetite was obtained at 5.5% (w/w) PS and 85% (w/w) moisture content at an initial pH of 4.5. However, the soil treated by PS showed a negative effect on soil health. The germination of mung bean (Vigna radiata (L.) R. Wilczek) and the CO2 release for the treated soil were low, indicating that toxicity had occurred in the treated soil. To restore the treated soil, adding a soil amendment, like CaCO3, fly ash, or crop residue, was able to improve plant growth and soil microbial activity.