Resumo. A same species can, under different environmental conditions, show variations in its growth trend and consequently in its final production. The dominant height (h100) is one of the most important variables for the forest growth and yield prediction, since we can see that a several number of yield functions are based on the dominant height description by site index equations. This paper had as objective to evaluate possible differences in the dominant height growth between two Eucalyptus saligna Smith clones and among different productivity soil classes, on which ones are implanted both Eucalyptus saligna clones. A set of permanent plots with annual or bi-annual measurements was used to fit different models of dominant height by age. Since the best model was defined and, utilizing the analysis of covariance, we tried to define the existence of different growth trends in dominant height. The growth trend analysis by using Backman’s function showed significant differences between both clones and among the different soil classes for clone 2864. Clone 4039 did not show a growth trend sharply different which would enough to separate this clone in two soil based groups. As a final result, we have that three different growth groups should be created to appropriately describe the dominant height growth: Group I – Clone 2864 and soil classes 1 and 2; Group II – Clone 2864 and soil classes 3, 4 and 5 and Group III – Clone 4039 and all soil classes.