Resumo. The prediction of a forest production requires knowledge of the dendrometric and edaphoclimatic variables. This study aimed to adjust and evaluate the accuracy of the 3-PG model in estimating the variables height, diameter at breast height (DBH), volume and mean annual increment (MAI) in commercial plantations of Eucalyptus spp. in northern of Minas Gerais state, using meteorological data from automatic surface stations and remote sensing. To validate the model, data of height, DBH, volume and MAI by 3-PG were estimated and compared with forest inventory data, totaling 335 measurements in 10 plots in the period 2005 to 2011. Linear statistical correlation methods (Pearson - r), the coefficient of determination (R²), the standard error of estimate (Syx), the standard error of the estimate in percentage (Syx%) and the efficient model (EF) were applied. It was observed that the model tends to overestimate the dendrometric variables when using the parameterization proposal for northern region of Minas Gerais state. After adjusting the parameters for the age of canopy closure (fullCanAge), the relationship between stem and DBH (nS) and parameterization of allometric equations for the total height and volume of wood incorporated in the 3-PG, obtained satisfactory results with values of R² = 0.94 and EF = 0.87 for DBH, height and volume and R² = 0.85 and EF = 0.58 for the MAI, as presented distribution of estimated data close to the average values of the observed data. The 3-PG was shown to be effective in estimating the DBH, height, volume and MAI become a potential tool for evaluating suitable areas for the expansion of new crops.