Abstract: Diamond detectors exhibit significant potential for measuring 14 MeV neutrons generated by fusion plasmas in tokamak experiments, owing to their exceptional radiation hardness and thermal stability. Monte Carlo simulation serves as a vital tool for designing and optimizing the performance of diamond detectors; however, discrepancies among the nuclear data libraries can considerably affect simulation outcomes. To systematically assess these influences, this study conducts comparative simulations based on five major international nuclear data libraries, with a focus on the response of diamond detectors to D-T neutrons under different nuclear data conditions. The results indicate a high level of consistency across libraries in both the ¹²C(n, el)¹²C and ¹²C(n, α)⁹Be reactions, with all libraries yielding a similar detection efficiency of approximately 0.035% for the latter. The main discrepancies emerge from the ¹²C(n, n+3α) reaction, where the CENDL library produces notably higher count rates in the corresponding energy region. Additionally, the response matrix of the diamond detector in the energy region of 1–20 MeV was simulated, offering essential input for forward simulations and spectrum unfolding. The sensitivity of the diamond detector to neutrons and γ rays across different energies was also investigated. This study provides a valuable theoretical basis and data support for the accurate simulation, appropriate nuclear data library selection, and structural optimization of diamond detectors in fusion neutron diagnostic applications.