Naturally-occurring iron minerals, goethite, magnetite, and hydrogen peroxide were used to catalyze and initiate Fenton-like oxidation of silica sand contaminated with mixture of diesel and kerosene in batch system. Optimal reaction conditions were investigated by varying pH(3, 7), $H_2O_2$ concentration(0%, 1%, 7%, 15%, 35%), initial contaminant concentration(0.2, 0.5, 1.0 g-mixture of diesel and kerosene/ kg-soil), and iron mineral contents(1, 5, and 10 wt % magnetite or goethite). Contaminant degradations in silica sand-iron mineral-$H_2O_2$ systems were identified by determining total petroleum hydrocarbon(TPH) concentration. The optimal pH of the system was 3. The system which iron minerals were the only iron source was more efficient than the system with $FeSO_4$ solution due to lower $H_2O_2$ consumption. In case of initial contaminant concentration of 1g-contaminant/kg-soil with 5 wt % magnetite, addition of 0%, 1%, 7%, 15%, and 35% of $H_2O_2$ showed 0%, 24.5%, 44%, 52%, and 70% of TPH reduction in 8 days, respectively. When the mineral contents were varied 0, 1, 5, and 10wt%, removal of contaminants were 0%, 33.5%, 50%, and 60% for magnetite and 0%, 29%, 41%, and 53% for goethite, respectively. Reaction of magnetite system showed higher degradation than that of goethite system due to dissolution of iron and mixed presence of iron(II) and iron(III); however, dissolved iron precipitated on the surface of iron mineral and seemed to cause reducing electron transfer activity on the surface and quenching $H_2O_2$. The system using goethite has better treatment efficiency due to less $H_2O_2$ consumption. When cach system was mixed by shaker, removal of contaminants increased by 41% for magnetite and 30% for goethite. Results of this study showed catalyzed $H_2O_2$ system made in-situ treatment of soil contaminated with petroleum possible without addition of iron source since natural soils generally contain iron minerals such as magnetite and goethite.