Abstract: Nuclear data as a fundamental input for reactor physics calculations, its sensitivities and uncertainties significantly impact core simulation accuracy. At present, some research results have been achieved in the research on the sensitivities and uncertainties of high-temperature gas-cooled reactor (HTR) in China, these research results mainly illustrated the need to pay special attention to the uncertainty caused by the elastic scattering cross-section of graphite in HTR. Gas-cooled Fast Reactor (GFR) usually use He and CO₂ as coolants, and almost no moderators, this is different from HTR. Furthermore, there has been no systematic work on GFR in China. It is necessary to conduct sensitivity and uncertainty research on GFR. In this paper, GFR such as ALLEGRO and GFR2400 were used as research objects. The sensitivity of k_eff to key nuclides such as Pu, U, Re, and W in ALLEGRO was calculated using the RMC code and the MCNP code respectively. Comparing the sensitivity calculation results of different programs, it is found that the relative deviations of the sensitivities of most nuclide-reactions calculated by RMC and MCNP are very small, but for the sensitivity of elastic scattering of some nuclides, the calculated results of the two codes are significantly different. Therefore, for nuclides with high sensitivity to elastic scattering cross-section, such as ²³⁸U (the relative deviation is −184.73%), ²³⁹Pu (the relative deviation is −20.60%) and ²⁴⁰Pu (the relative deviation is −50.32%), the sensitivities were re-performed using the direct perturbation method. The sensitivity calculations using the direct perturbation method show that the calculation results of MCNP are more accurate. The calculation method of RMC to calculate the sensitivity of elastic scattering cross-sections in GFR may need to be further improved, which verifies the sensitivity calculation ability of RMC in GFR. The RMC was used to calculate the sensitivities and uncertainties of key nuclides such as Pu, U, Re, and W in GFR2400, and compared with the results in the public literature. It is found that due to the differences in models, the sensitivity calculation results of some nuclides-reaction have significant deviations, but the uncertainty calculation results are in good agreement, which verifies the uncertainty calculation ability of the RMC in GFR. The results of this paper verify the application of the RMC in calculating nuclear data sensitivity and uncertainty, and can provide support for GFR design and nuclear data research.