This study aimed to compare two diameter distribution models in Eucalyptus grandis x Eucalyptus urophylla submitted to thinning; and evaluate the influence of two adjustment methods of probability density functions on the quality of the diameter structure projections. The data used in this study came from an experiment on thinning of eucalyptus clones, located in the northeast region of the state of Bahia. The measurements were made at the ages 27, 40, 50, 58, 61, 76, 87, 101, 112, 122, 137, 147, 158 and 165 months, with thinnings carried out at 58 and 142 months. Two diameter distribution models were used to prognosticate the diametric structures of the stands, starting from an observed distribution at ages 27 (before first thinning), 61 (after first thinning) and 147 (after second thinning) months. Three systems were obtained according to the adjusting method of the Weibull function: System 1, using the linear approximation; Systems 2 and 3, using the maximum likelihood. The goodness of fit of prognosticated structures in relation to those observed was verified by the Kolmogorov-Smirnov test at 5% of significance. System 2 was the most suitable for practical applications for thinning in eucalypts. The maximum likelihood method should preferably be used when compared to linear approximation, ensuring a better quality in the prognoses of the diameter distribution. © 2020 University of Sao Paolo. All rights reserved.
