The goal of the research was to investigate the retention capacity of six green roof models (SHP1, SHP2, SHP3, SH, S, and SP) constructed with the use of the square-shaped plastic trays, Floradrain FD 25 drainage elements, SF filter sheets, and the specified extensive substrates (with or without the hydrogel amendment). The SHP1 and SHP2 models were constructed in March 2017, SHP3 and SHin November 2017, while S and SPin April 2018. Four models (SHP1, SHP2, SHP3, and SP) contained the plants (the goldmoss stonecrop Sedum Acre), whereas two models (S and SH) did not contain the vegetation. The substrates of SHP1, SHP2, SHP3, and SH models contained the hydrogel admixtures. The investigations were conducted with the use of simulated (and partially natural) precipitations. The water retention capacity of each green roof model was established based on the difference between the precipitation volume and the volume of runoff from a model. The results show that green roofs can be useful stormwater management tools. The calculated stormwater retention rates ranged from 29.50% to 85.15%. In most cases, the best water retention capacity was exhibited by the SHP3 model, constructed in November 2017 and planted in April 2018, containing the substrate amended with superabsorbent (cross-linked potassium polyacrylate). The similarly constructed SHP1 and SHP2 models, which were built in March 2017, in some cases had lower water retention capacity. These models contained older hydrogel and were overgrown with older, smaller, and worse looking plants, partially supplanted by mosses. Such results indicate that the efficiency of hydrogel may decrease over time. In many cases, the S (not vegetated, without hydrogel), SH (not vegetated, with substrate containing hydrogel), and SP (vegetated, without hydrogel) models had slightly lower water retention capacity. The results of investigations indicate that there was a relatively strong positive linear correlation between the retention depth and duration of the antecedent period elapsed from the preceding total (or substantial) saturation of the green roof models (labelled in this article as period since total saturation-PSTS). The weather conditions i.e. air temperature and relative humidity as well as PSTS are very important parameters that influence the retention capacity of the green roof models. The result show that duration of PSTS can be stronger correlated with the retention depth than antecedent dry period (ADP) elapsed from the end of last precipitation, regardless of its depth and intensity.