Abstract Representative experimental results for the heat transfer coefficients (HTCs) of liquid metals flowing in tubes under turbulent conditions are reviewed from the standpoint of experimental apparatuses, instrumentations and data derivation methods. Fourteen experiments are investigated together with replicated schematics of test sections or apparatuses in literatures. These HTC data are classified into four groups based on the method to determine three important elements to derive the heat transfer coefficient. The three important elements are heat flux, wall surface temperature and bulk temperature. Group-1 data use measured wall temperatures and bulk temperatures derived on the basis of the measured temperature and velocity profiles in the fluid. These data trends located on the higher Nusselt number sides among the reviewed data. While, in group-2 data, the bulk temperature is linearly interpolated from the measured inlet and outlet temperatures. The data trends of this group show lower in some extent than the former group especially in the low Peclet number region. It was clarified from the schematic of the test sections and the experimental methods that some measured data contain unclear factors. These data trends classified in the group-3 are clearly under the former two trends. The CFD simulation results agree with the group-1 data. The impact of using a simply interpolated bulk temperature that causes the HTC to lower is evaluated, and it is concluded that the impact can be neglected under the Peclet number larger than 1000.