There is strong interest in being able to predict the air thermal distribution in data centers to have a better understanding of the problem so that cooling energy can be reduced. As experimental studies in data centers require exhaustive measurement efforts, CFD presents an alternative solution. A few of these studies were validated against experimental data, and relatively large errors were found between numerical results and experimental data. These errors could be due to either a non-controlled experiment (e.g. floor flow leakage and leakage path on servers' racks) or non-suitable boundary conditions (e.g. thermal boundary conditions on the floor) in the numerical simulations. Another important factor that is commonly ignored in data center CFD analysis is the modeling of the perforated surfaces (e.g. tile or rack back door) with the correct mass and momentum of the airflow. The study of Abdelmaksoud et al. [1, 2] shows the importance of including the correct momentum issuing from perforated tile in data center CFD simulations. In the present paper, we have applied various perforated tile models on a small data center test cell that consists of three high-power simulated racks generating up to 100kW heat load in total. Including these perforated tile models in the CFD analysis show a significant effect on the CFD results. Further discussion about modeling techniques of the perforated tile in the data center CFD simulations is presented in this paper.