Abstract: The nuclear data are one of the most important uncertainty sources for the reactor physics calculation. The quantification of the uncertainty introduced by the nuclear data is an important aspect in reactor uncertainty analysis. On the other hand, it is significant to improve the accuracy of the nuclear data for the reliability of reactor calculation results as well as the economy and safety of the reactor. The target accuracy assessment is one of the most effective approaches to give the accuracy requirement of the nuclear data to reduce calculation result uncertainties based on the sensitivity and uncertainty analysis. In this work, a two-step approach was proposed for sensitivity calculation. The k_eff sensitivity and uncertainty analysis was performed based on the fast reactor benchmark BN-600. The optimization mathematical model related to the target accuracy assessment was built and the differential evolution algorithm was applied to solve the optimization problems. The uncertainty requirements of the nuclear data were given when the target accuracy of k_eff was set to 0.3%.