distributed radar objects are represented as a set of a large number of point scatterers. Such models are called multipoint models. They are used for mathematical and simulation modeling of the angular glint of distributed targets consisting in the control of its precision reflection altimeter (PRA) formation and correlation function (Guo et al., 2018). Multipoint models replace the whole distributed radar target with high accuracy considering the shape, movement and Doppler spectrum parameters. The number of scatterers can reach hundreds or more. This makes it difficult to use these models in the semi-realistic modeling or in any real device. As a measure of reducing the number of scatterers (model points), the transition to the so-called low-point geometric models is used (Artyushenko & Kiselev, 2015;Stepanov & Kiselev, 2019). They contain much fewer points, but can be applied for replacing only part of the radar target lying inside of the one radar resolution cell at the same time. The low-point model synthesis is based on multipoint model in purpose of providing the geometrical and Doppler spectrum parameters of the replacing radar target with the high accuracy. Thus, for objects distributed over two angle coordinates, five or nine points (Artyushenko & Kiselev, 2015) are sufficient, which emit uncorrelated normal random processes with proper correlation properties (R i (τ) and S i (τ)-correlation functions of the same and dissimilar quadrature components, the signal emitted from the ith point of the model). Models whose points emit uncorrelated signals are called incoherent geometric models (Artyushenko & Kiselev, 2015). The next step toward reducing the number of points is the transition to partially coherent models, in which radiations from the points are statistically related and correlated. A two-dimensional partially coherent geometric model containing only four points located in the vertices of a square is known (Artyushenko et al., 2018). Correlated normal random processes with the given proper correlation properties (B i (τ) and C i (τ)-correlation functions of the similar and dissimilar quadrature components, the signal emitted from the ith point of the model) are brought to the points of such a model.In Stepanov and Kiselev (2019), the possibility of transition for one-dimensional incoherent models to partially coherent ones is proved based on the equivalence of the two-point partially coherent and three-point incoherent model. The equivalence of these models includes the providing equivalent PRA formations and correlation functions of angular glint. With this transition, the number of model points is reduced from three to two. The results obtained in Stepanov and Kiselev (2019) allow determining the parameters of partially coherent models (R i (τ) and B i (τ)) from the known parameters of one-dimensional incoherent models (mutual correlation coefficients of signals and their S i (τ) and C i (τ)). This work aims to develop these results to a level that allows the transition from a two-dimensional g...