Abstract: The development of high precision burnup analysis requires high accuracy burnup database. The benchmarking method of burnup database is very important for developing high accuracy burnup database. Based on the burnup database composed of the multitemperature continuous point section database CENACE1.3 made by China Nuclear Data Center and Cinder90 burnup database, the modeling of SF95 samples in the post irradiation experiment of TAKAHAMA3 power water reactor was taken as an example. The spent fuel composition database SFCOMPO2.0 gives the detailed information of postirradiation experiment of TAKAHAMA3 power water reactor, including component information, power history, boron concentration history and nuclides inventory value. The influence of modeling elements on burnup calculation was studied, and the modeling method of burnup benchmark experiment was determined. The influence of modeling elements such as material temperature and density, boron concentration and burnup step on modeling calculation was determined. As an important part of burnup credit, burnup analysis aims to obtain the nuclide composition and inventory of spent fuel. Using the burnup benchmark experimental modeling method, the burnup credit was studied and the comparison of calculated values and experimental values of interested actinide nuclides and fission products nuclides was carried out. The preliminary results showed that the deviation between the calculated values and the experimental values of main actinide nuclides is less than 2%, the relative deviation between the calculated values and the experimental values of most minor actinides nuclides is less than 10%, and the deviation between the calculated values and the experimental values of most important fission products nuclides is less than 5%. In this paper, based on the adjacent burnup chain of 125Sb, the variation law of 125Sb inventory with burnup depth was theoretically analyzed, and it is confirmed that there are defects in the measurement results of destructive radiochemical experiment. In this study, the destructive experimental value of 125Sb was modified by nondestructive experimental measurement value and experimental value modification formula. And the corrected experimental values of 125Sb inventory were obtained, which reduced the calculation deviation between calculated value and experimental value from nearly 170% to less than 20%. This study shows that the benchmarking of burnup database using the spent fuel composition database requires not only the appropriate burnup benchmark experimental modeling method, but also the appropriate evaluation and verification of inventory measurement data, so as to be able to be used for benchmarking of burnup database and burnup credit research.