Predictions for confined diffusion flames are made by using density-unweighted and density-weighted averaging approaches. Predicted results using these two mathematical models are validated by comparing them with the available experimental data in open literature. Significance of dilatation effect is also demonstrated in the study. Prediction/measurement comparisons reveal that the performance of the density-weighted model is unsatisfactory, although the density-weighted averaging approach is preferred for the exothermically reacting flows in which the density plays an important and pervasive role in the dynamics of turbulence. More testing needs to be done to establish the empirical constants used for the density-weighted model. In addition, new benchmark experiments on diffusion flames should be conducted to provide a well-defined data base, including complete inlet boundary conditions, for further work.