In this project a fast, easy and cost effective method base on the use of magnetic montmorillonite was applied for the extraction and determination of vitamin B 12 in food stuffs. The “vitamin B 12 ” refers to a group of chemical compounds of similar structure that play an important role in the formation of red blood cells, functioning of the nervous system and translocation of the methylene group in DNA synthesis. In addition, it has an importantfunction in human physiology. Vitamin B 12 deficiency inhumans is manifested by anemia and neuropathy. In this method, the extracted analyte was determined by high performance liquid chromatography (HPLC) and visible detection. In the proposed method, Fe 3 O 4 /montmorillonite nanocomposite (Fe 3 O 4 /MMT NC) was easily synthesized by mixing of sodium montmorillonite, Fe 2+ and Fe 3+ in the hot water (80?C) at the nitrogen atmosphere. The prepared magnetic nanocomposite was used as the extracting substrate. In this procedure, vitamin B 12 was extracted from 3ml of aqueous solution containing 1mg extracting nanocomposite. After the extraction, the extracting sorbent containing the target analyte was collected by the means of an external magnet and the supernatant was decanted. The extracted analyte was desorbed by 50µL of methanol as desorption solvent. After that, 40 µL of desorption solvent was injected to HPLC. The extraction parameters such as sorbentamount, sample solution pH, type and volume of desorption solvent,extraction and desorption time were investigated and optimized. The optimum condition was sorbent amount of 1mg, sample solution pH=5, 50 µL methanol as desorption solvent, 5 min extraction time and 2 min desorption time. Under the optimum condition, linear dynamic range (LDR) was obtained from 0.02-1 ?g/L and the correlation coefficient (r 2 ) was 0.9933. The limit of detection (LOD) and limit of quantitation (LOQ) was determined 8.5 ?g/L and 16.8 ?g/L, respectively. Relative standard deviation (RSD) for the concentration of 50 µL and 200 µL was 2.2% and 3.2%, respectively. To investigate the capability of the method in analysis of real samples, some food samples such as liver, meat, beef and milk samples was analyzed by the method. The recoveries were determined between 66.1% and 81.4%.