Abstract. Constructed Technosols are important means to substitute natural soil material such as peat and geogenic material to be used in urban green infrastructure. One of the most important features of such soils is related to the water cycle and can be described by the soil water retention curve (WRC). The WRC depends on the composition of the constructed Technosols e.g. their components and their mixing ratio. The diversity of possible components and the infinite number of mixing ratios practically prohibit the experimental identification of the optimal composition regarding the targeted soil functions. In this study we propose a compositional model for predicting the WRC of any binary mixture based on the measured WRCs of it’s two pure components only (basic scheme) or with one additional mixture (extended scheme). The model is developed from existing methods for estimating the porosity in binary mixtures. The compositional model approach was tested for four data sets of measured WRCs for different binary mixtures taken from the literature. To assess the suitability of these mixtures for typical urban applications, the distribution of water and air in 50 cm high containers filled with the mixtures was predicted under hydrostatic conditions. The difference between the maxima of the pore-size distributions ∆PSDmax of the components indicates the applicability of the compositional approach. For binary mixtures with small ∆PSDmax, the water content deviations between the predicted and the measured WRCs range from 0.004 to 0.039 m3 m−3. For mixtures with a large ∆PSDmax, the compositional model is not applicable. The knowledge of the WRC of any mixing ratio enables the quick choice of a composition, which suits the targeted application.