The paper reveals regularities of grass projective cover variation depending on the level of recreational transformation of public green spaces, taking into account stand density and soil electrical conductivity. The hypothesis that herbaceous cover is positively influenced by moisture and trophicity of edaphotopes and the stand is an antagonist of herbaceous cover is confirmed. The nature of intra-ecosystem relationships of the park stand is influenced by the level of recreation. Assessment of the level of recreational transformation of the grass cover of green areas of public use is performed on the example of the park Novoalexandrovsky of Melitopol. Within the city park, the grass layer is influenced by natural and anthropogenic environmental factors. Based on the data on the hardness of the soil quantified the level of recreational load and set the following levels of recreational load: low level, below a moderate level, moderate level, above a moderate level and high level. Stand density is an important natural factor that affects the herb layer primarily through the regulatory function of access to sunlight. Quantitatively, trophicity and moisture can be estimated using the electrical conductivity index, which depends on both the amount of salts dissolved in the ground liquid (trophicity) and the moisture content. The obtained data show that electrical conductivity of soil within the park is on the average 0,11 ± 0,0037 dSm/m and in 95% of cases varies from 0 (electrical conductivity below the lower limit of device sensitivity) to 0,19 dSm/m. Observed electrical conductivity values are much less than critical values, so within the park, excessive salinity is not a limiting factor for herbaceous plant growth. Projective cover of the herbaceous layer is on average 2,60 ± 0,11%. In 95% of cases this indicator is in the range of 0‒10%. The level of variability of projective coverage is rather high ‒ 96,48%. Distribution asymmetry is significant and positive, indicating a significant shift in the distribution of this indicator towards small values. The kurtosis is also significant and positive, which indicates the predominance of modal values compared to the random alternative. The distribution of the projective coverage can be described as a mixture of two distributions with a normal law. The component with a lower mean of 0,76 ± 0,59% is 46,0% of the mixture, and the component with a mean of 3,63 ± 2,24% is 54,0% of the mixture. It should be noted that the Kolmogorov‒Smirnov test indicates a low level of explanatory power even for such a complex model. Therefore, it should be assumed that the variation of the projective cover of herbaceous cover is determined by a number of significant factors. According to the quantitative indicators of vegetation cover, the following levels of recreational load can be distinguished: low (combines small and less than moderate levels), medium (combines moderate and above moderate) and high (corresponds to the high level). Differences in projective cover within the smallest and below moderate level on the one hand, and medium and above moderate level on the other hand, are not statistically reliable.