As small modular reactors continue to advance, heat pipe reactors (HPRs) are being developed by multiple organizations. Designing heat pipes with optimal heat transfer capability is crucial for the effective performance of HPRs. Based on NSGA-III genetic algorithm, this paper presented the design method of a high-temperature heat pipe featuring a composite wick. The study examined the influence of heat pipe structure parameters on performance parameters and heat transfer limits. Based on the analysis, we built multi-objective modeling of heat pipe and obtained the optimal structural parameters by NSGA-III. Verification confirmed that the heat pipe met the Mach number and effective capillary radius requirements. This study provided valuable insights for improving heat pipe manufacturing techniques and reactor miniaturization.