Abstract: The spacer grid models are important models for simulating the thermal-hydraulics phenomena during reflooding process. The spacer grid pressure drop model, wall heat transfer enhancement model, interfacial heat transfer enhancement model, droplet breakup model, spacer grid re-wet model and grid temperature model are integrated in COSINE multiphase sub-channel code. In this paper, the typical rod bundle heat transfer (RBHT) reflooding tests were selected to assess the spacer grid model in the code. The calculation results show that the spacer grid models can significantly enhance the code predicted heat transfer capacity, accelerate the quench rate and reduce the peak cladding temperature (PCT) during the reflooding process, and the calculated results are in good agreement with the experimental measurements. In addition, the deviation of the PCT predicted by the code is reduced from 10% to 7% due to the adoption of the spacer grid models. The spacer grid models have the ability to enhance the interfacial heat transfer between vapor and droplet, therefore the calculated results with the spacer grid models show a significantly faster decrease in vapor temperature than that without the spacer grid models. And the calculated vapor temperatures with spacer grid models are in good agreement with the experimental measurements. The trend of the code predicted spacer grid temperature agrees well with the experimental results, but the code predicted spacer grid temperature is slightly higher than the experimental measurements. Furthermore, the early wetting of the temperature probe located on the spacer grid in the experiment results in a rapid decrease in the spacer grid temperature, however，the code predicted spacer grid temperature does not show as sharp a decrease as the experimental results due to the interpolation method in the calculation of the spacer grid temperature. The spacer grid models adopted in the code can simulate the breakup of droplets across the spacer grid, the diameter ratio of the shattered droplets predicted by the code is in good agreement with the experimental measurements, and the calculation results of the Cheung model are more accurate than those of the Paik model in the test case 7151. In addition to this, the code can predict the trend of droplet velocity near the spacer grid during reflooding process, but the calculated droplet velocity decreases significantly faster than the experimental measurements, and it needs more work in future. In general, the calculated results with the spacer grid models are in good agreement with the experimental results, and the spacer grid models adopted in the code are reasonable and reliable.