Cenotic diversity and leading ecological factors of its floristic differentiation were studied on an example of two areas – Kyiv parks "Nivki" and "Teremki". It is shown that in megalopolis the Galeobdoloni-Carpinetum impatientosum parviflorae subassociation is formed under anthropogenic pressure on the typical ecotope of near-Dnieper hornbeam oak forests on fresh gray-forest soils. The degree of anthropogenic transformation of cenofloras can be estimated by the number of species of Robinietea and Galio-Urticetea classes, as well as neophytes and cultivars. Phytoindication for hemeroby index may be also used in calculation. We propose the modified index of biotic dispersion (normalized by alpha-diversity) for the estimation of ecophytocenotic range (beta-diversity) of releves series. We found that alpha-diversity initially increases (due to the invasion of antropophytes) at low level of antropogenic pressure, then it decreases (due to the loss of aboriginal species) secondarily with increasing of human impact. Also we found that beta-diversity (differential diversity) decreases, increasing homogeneity of plant cover, under the influence of anthropogenic factor. Vegetation classification was completed by a new original method of cluster analysis, designated as DRSA («distance-ranked sorting assembling»). The classification quality is suggested to be validated on the "seriation" diagram, which is а distance matrix between objects with gradient filling. Dark diagonal blocks confirm clusters’ density (intracluster compactness), uncolored off-diagonal blocks are evidence in favor of clusters’ isolation (intercluster distinctness). In addition, distinction of clusters (syntaxa) in ordination area suggests their independence. For phytoindication we propose to include only species with more than 10% constancy. Furthermore, for the description of syntaxonomic amplitude we suggest to use 25%-75% interquartile scope instead of mean and standard deviation. It is shown that comparative analysis of syntaxa for each ecofactor is convenient to carry out by using violin (bulb) plots. A new approach to the phytoindication of syntaxa, designated as R-phytoindication, was proposed for our study. In this case, the ecofactor values, calculated for individual releves, are not taken into account, however, the composition of cenoflora with species constancies is used that helps us to minimize for phytoindication the influence of non-typical species. We suggested a syntaxon’s amplitude to be described by more robust statistics: for the optimum of amplitude (central tendency) – by a median (instead of arithmetic mean), and for the range of tolerance – by an interquartile scope (instead of standard deviation). We assesses amplitudes of syntaxa by phytoindication method for moisture (Hd), acidity (Rc), soil nitrogen content (Nt), wetting variability (vHd), light regime (Lc), salt regime (Sl). We revealed no significant differences on these ecofactors among ecotopes of our syntaxa, that proved the variant syntaxonomic rank for all syntaxa. We found that the core of species composition of our phytocenoses consists of plants with moderate requirements for moisture, soil nitrogen, light and salt regime. We prove that the leading factor of syntaxonomic differentiation is hidden anthropogenic, which is not subject to direct measurement. But we detect that hidden factor of "human pressure" was correlated with phytoindication parameters (variables) that can be measured "directly" by species composition of plant communities. The most correlated factors were ecofactors of soil nitrogen, wetting variability, light regime and hemeroby. The last one is the most indicative empirically for the assessment of "human impact". We establish that there is a concept of «hemeroby of phytocenosis» (tolerance to human impact), which can be calculated approximately as the mean or the median of hemeroby scores of individual species which are present in it.