This paper presents a study of uncertainty quantification (UQ) of spent nuclear fuel isotope inventory. Takahama-3 realistic assays were used for calculation to quantify the impact of three different parameter groups, namely, design parameters (i.e., fuel pellet radius, clad outer radius, UO2 density, and UO2 enrichment), operation conditions (i.e., power density, fuel temperature, moderator temperature, and boron concentration), and covariance of nuclear data. The UQ study is focused on analysis of fission products and actinides used in criticality analysis during the back-end cycle. Stochastic sampling method was used for the UQ and Lagrange interpolation is used for the prediction of isotope inventory. In-house two step code system, STREAM/RAST-K was used for calculation with ENDF/B-VII.1 cross section library and ENDF/B-VII.0 decay library. Most of the actinides and fission products are calculated within +/- 5% of experiment. In addition, U-238 and Pu-239 have the largest contribution on k(eff) uncertainty. Moreover, Cm-245 and Cm-246 have the largest isotope inventory uncertainty compared with other isotopes and the covariance of nuclear data has the most dominant effect on the uncertainty of isotope composition. (C) 2021 Elsevier Ltd. All rights reserved.