Background and Aims: In mountain ecosystems, predicting root density is highly challenging due to the spatial heterogeneity of forest communities. We present a simple and semi-mechanistic model, named "ChaMRoots," that allows the prediction of root interception density (RID, number of roots/m²) by hypothesising that RID at a given point is contributed by the roots of the surrounding trees. Methods: The model comprises three sub-models for predicting the: (i) spatial heterogeneity: RID of roots of ]0, 1] mm in diameter within the top soil layer of 0.2 m as a function of tree basal area at breast height, and the distance between the tree and a given point; (ii) diameter spectrum: the distribution of RID as a function of root diameter up to 50 mm thick; (iii) vertical profile: the distribution of RID as a function of soil depth; The fitted RID data were measured at two uneven-aged mountain forest ecosystems in the French Alps. These sites differ in tree density and species composition. Root density in 25 mini-monoliths (0.3 m deep) and 32 soil profiles (1.0 m deep) were measured, along with the surrounding trees' dimensions and positions. Key Results: The model achieved a good compromise between the number of aerial input parameters and the fit to the observed data. Conclusions: ChaMRoots provides a facilitated and highly transplantable approach for modelling three-dimensional root spatial distribution in complex forest ecosystems