Fluoride salt-cooled high-temperature reactor (FHR) is one of the Gen IV nuclear systems. It utilizes small spherical type fuel with 30mm diameter, and the core is filled with numerous pebbles. The authors developed a simple code that generates randomly-packed spherical fuels inside the cylindrical core. The fluid domain was generated and converted to perform the Computational Fluid Dynamics (CFD) analysis to figure out the local heat transfer coefficient of pebble-bed fuels. To ensure promising CFD analysis model the authors examined sensitive parameters – the number of pebbles, grid size and gap size and turbulence models. Large-Eddy Simulation (LES) was performed for the selection of the turbulence model with face-centered cubic (FCC) single channel model. As a result, k-omega Shear Stress Transport (SST) with gamma transition model is selected as a turbulence model for randomly-packed pebble’s CFD analysis. The result shows that the pebble’s local heat transfer coefficient has a Gaussian distribution with average and standard deviation. Furthermore, the authors propose a new Nusselt number correlation for the randomly-packed pebble bed reactor with FLiBe coolant. The result leads to a conclusion that the thermal-hydraulic performance of fuel has a statistical distribution and it will have the effect on the robustness of fuel material and design criteria of safety systems.