Abstract: Due to excellent thermal conductivity, low expansion, irradiation resistance and good helium brittleness, ferrite/martensite (FM) steel and its oxide diffusion enhancement (ODS) steel are candidates for future fission reactor envelope and fusion reactor. By replacing Mo, Nb and Ni with elements such as W, V and Ta, the reduced-activation ferrite/martensitic (RAFM) steel developed on the basis of FM steel can further reduce the long-life radioisotopes after reactor use. In the reactor, the structural materials suffer not only high temperature and high pressure working environment, but also high flux neutron irradiation. Irradiation usually induces or promotes the segregation of alloy elements at the grain boundary, and the segregation of different elements on the grain boundary will enhance or weaken the strength of the grain boundary, and the decrease of the grain boundary strength will cause the embrittlement of materials. Therefore, the study of the grain boundary segregation behavior of solute elements is significant to the development and performance prediction of RAFM steels. In this paper, the large atomic/molecular massive parallel simulation program (LAMMPS) developed by Sandia National Laboratory was used to study the segregation behavior of Cr/W elements at grain boundaries in different model alloys (FeW and FeCrW alloys) at 300-800 K, and effects of vacancies and temperature on alloy elements, and the coupling of W and Cr elements were discussed as well by the molecular dynamics and Metropolis Monte Carlo methods. The calculation results show that W element will be significantly enriched at grain boundary in FeW alloy. W is not observed to be enriched, while the distribution behavior of Cr element has two cases: grain boundary segregation and group segregation in the bulk phase in FeCrW alloy. Such calculations suggest that the addition of Cr element will inhibit the enrichment of W at the grain boundary. Combined with some subsequent computational simulation results and research results of other scholars, it is believed that the main reason is the repulsion between Cr and W. In addition, the calculation results suggest that vacancies have no significant effect on the segregation behavior of W at grain boundaries, but the vacancy and temperature have a synergistic effect on the distribution of Cr. Last but not the least, with the increase of temperature, the solubility of solute elements in the body phase increases, and the degree of segregation at the grain boundary gradually decreases, that is, the temperature increase slightly limits the grain boundary segregation behavior of Cr and W.