Plant root system plays an essential role in the acquisition of the edaphic resources, which are subject to local depletion. The size as well as the architecture of the root system determines the efficiency of the acquisition. In the present study, a stochastic model of plant root system architecture is formulated. The continuous growth and development of root system is described and modelled by stochastic processes (discrete events associated with a certain probability). The parameters of the model for each growth cycle include branching probability, w (rhythm ratio main axis vs. lateral roots), b (probability of growth) and c (probability of survival). Root segments were presented as connections of individual nodes. As root has no nodes in the sense of the botanical terms, an imaginary node with an elementary length is introduced. In order to obtain the parameters of the model, winter wheat seedlings were grown in a phytotron in sand culture watered by nutrient solution. Individual roots of 19-days-old se dlings were scanned and the images obtained were analysed with a root image-analysing software WinRhizo. Roots were clustered into 3 relatively homogeneous groups after an analysis of similarity according to 4 criteria: length of main axe, diameter of root apex of the main axe, lateral length density (total length of lateral roots per unit of main axe length), lateral root density (number of lateral roots per unit of main axe). In each root group, the parameters were fitted with a non-linear generalised least square method by comparing the theoretical length of root segments of various orders with the experimental data.