Resumo. This study was conducted in a semideciduous forest fragment of 52 hectares, known as Rosal Forest, in the municipality of Guaçuí, Espírito Santo state. It had the objective to test vertical stratification methodologies, evaluating the appropriateness of these methodologies based on the ecological successional classification of the species. To do so, the method of sampling fixed area was used, and distributed 16 plots of 600 m2, in a systematic way. All individuals with DBH ≥ 5 cm, had their total height measures directly. The trees went through the botanical classification identification process and had their classifications in accordance with their ecological groups. The analysis of the tree vertical structure was made by the distribution of the number of trees in the strata, using four methods: Method 1 - Souza (1990), Method 2 - Souza et al. (2003), Method 3 - IUFRO, and Method 4 - Calegário et al. (1994). The classification results were compared to the successional classification of species, according to the ecologic groups that they belong to. The method 1 did not allow a detailed analysis about the behavior of stratum II species, because of the strong tendencies to concentrate a greater number of individuals in this stratum, already the Method 2 has allowed a greater detail. The Method 3 minimized the problem encountered in Method 1, but the dominant height change of the sample can change the whole discussion, demonstrating the method weakness. Method 4 did not bring good results for the forest stratification under study, because this method only divided in two height strata.