Abstract: The Nuclear Hot Spring (NHS) is a molten salt pebble-bed reactor featured by full power natural circulation. The unique horizontal coolant flow of the NHS demands the fuel recycling schemes based on radial zoning refueling and the corresponding method of fuel management optimization. The local searching algorithm (LSA) and the simulated annealing algorithm (SAA), the stochastic optimization methods widely used in the refueling optimization problems in LWRs, were applied to the analysis of refueling optimization of the NHS. The analysis results indicate that, compared with the LSA, the SAA can survive the traps of local optimized solutions and reach the global optimized solution, and the quality of optimization of the SAA is independent of the choice of the initial solution. The optimization result gives excellent effects on the in-core power flattening and the suppression of fuel center temperature. For the one dimensional zoning refueling schemes of the NHS, the SAA is an appropriate optimization method.