This study aimed to test different statistical models to adjust site index curves and apply models of identity tests in comparing the mean curves of dominant height in Pinus taeda to two regions of the West of the State of Santa Catarina, Brazil. Data dominant heights were obtained from permanent plots, with 600 m² size, distributed randomly. Selecting the best adjusted statistical model was made through the following adjustment and accuracy of statistics: determination coefficient (R²adj), standard error of estimate (Syx) in absolute and relative terms and graphic distribution of waste. The test Regazzi (2003) was applied to analyze the possibility of using a single model of the dominant height for the two regions. The Prodan model was chosen for construction of site index curves by the method of the guide curve. By applying the test Regazzi (2003), a single equation is not sufficient to represent the growth of dominant height in the two regions.