Polymer flooding of heavy oils on the laboratory scale shows appreciable incremental tertiary oil recovery. In reality, however, this high recovery efficiency usually cannot be achieved in the field due mainly to extremely unfavorable mobility ratio and reservoir heterogeneity. The former promotes viscous fingering while the latter induces channeling; hence both of these factors make the displacement process less efficient. This paper identifies the dominant scaling groups for polymer flooding currently conducted in western Canadian heavy oil reservoirs. Twenty-eight dimensionless scaling groups governing the process of polymer flooding for enhanced heavy oil recovery were derived using inspectional analysis, and a fully tuned numerical model for polymer flooding of a heavy oil sample in a two-dimensional sand pack was then developed to validate the effectiveness of these scaling groups. A good agreement among different cases with the same group values was observed, showing the validity of the scaling groups. The effect of each scaling group on oil recovery was examined by numerical sensitivity analysis. By doing so, nine scaling groups dominating polymer flooding enhanced heavy oil recovery were identified. These dominant scaling groups can be used to design scaled experiments to predict field-scale oil recovery by polymer flooding in heavy oil reservoirs.