In the present work the temperature fluctuations in a mixing tee were simulated on FLUENT platform using the Large-eddy simulation (LES) turbulent flow model with the sub-grid scale (SGS) model of Smagorinsky-Lilly (SL). The temperature and velocity fields, the normalized mean and fluctuating temperatures and velocities were predicted and analyzed with consideration of buoyancy. The normalized mean and fluctuating temperatures were defined to describe the time-averaged temperature and the time-averaged temperature fluctuation intensity. The numerical results of the normalized mean and fluctuating temperatures were compared with those of the experimental ones published in previous literature, which shows that numerical results have good agreement with the experimental data. The temperature fluctuations and power spectrum densities (PSD) at the locations having the strongest temperature fluctuations both in tee junction and on the walls were analyzed to evaluate the potential of thermal fatigue. The LES flow simulation and power spectral analysis are helpful for the Integrity evaluation of the structures such as the tee junction, elbow, piping system to predict the temperature fluctuation and thermal stripping in a tee junction of mixing hot and cold fluids.