Even in the absence of flow barriers and other reservoir heterogeneities, current SAGD models underestimate heavy oil recovery, and are generally seen to be inaccurate in predicting important production characteristics like steam oil ratio (SOR) and water oil ratio (WOR), when compared against experimental or field data. It has been hypothesized by some researchers that water-in-oil emulsions are formed at the steam-oil interface. It is our premise that transport of these emulsion droplets into the bitumen phase facilitates convective heat transfer resulting in improved recovery. Incorporating these effects is key to accurately modeling the SAGD process. Unfortunately, the physics of emulsion formation and transport in porous media is not well understood, and current simulators do not have the capability to directly model such effects. A new approach that models the emulsion droplets as chemical species and accounts for the dispersion and adsorption phenomena is implemented in this paper. This model utilizes the features available in most commercial simulators in order to model emulsion generation propagation and coalescence in porous media. The results from such a mechanistic simulation are compared against published SAGD experimental data. Our results show significant improvement from previous SAGD models and bolster the argument that emulsions are responsible for a key heat transport mechanism during SAGD. The simulation is also being used currently to design laboratory experiments to verify the formation of emulsions and their net effect on heat transfer.