The article discusses a new approach to the creation of complex models — two-dimensional geometric objects. The basis of this approach is the topological characteristics of the model, which are set by the user in separate selected nodes — cluster points, which provides the necessary flexibility of description and allows changing the external representation of the object to a considerable extent. The cluster representation has a sufficient degree of universality, allowing the formation of rather complex geometric models. At the same time, an important feature of such a parametric representation is the ability to add necessary geometric characteristics by simple means and modify existing ones. The basic concept of a cluster point introduced in the paper is based on the possibility of specifying a set of parameters for a local description of a closed geometric region surrounding a given point. Combining a finite set of these areas allows you to create a new parameterized geometric object as a whole, which was called the MUC object (an object built on the model of a unified cluster). Drawing an analogy with computer graphics, we can say that the MUC-object is a vector image of a complex geometric shape (as opposed to a raster image, which is defined by the color of each pixel on the screen). In other words, the proposed geometric description method allows you to convert a complex geometric image into a "vector" format, and thus significantly reduce the total amount of stored data (as opposed to a pixel image). The article describes a universal set of parameters relating to a separate cluster point, which allows one to construct both the simplest geometric objects and more complex ones, based on predetermined algorithms for combining them. The presented parameters allow us to construct the boundary of a geometric object in the form of a piecewise-smooth line, the form of which depends on the specified angles and types of connecting lines at each cluster point. The possible combinations of lines for the IOC object, consisting of one cluster point, as well as the combination of two and three points are listed. For the last two cases, merging algorithms are described, also based on the parameters specified at cluster points. The paper concludes that a cluster object consisting of a single point is a generalization of classical geometric figures — a square, a rectangle, a circle and an ellipse and allows considering the algorithm of their construction from a unified position. At the end of the article, the software implementation of the presented approach is briefly described, which allows us to construct not only the above-named figures, but also, as an example, a five-element MUC object. In the latter case, there are several screen shots of this object with modified parameters, which demonstrates that, despite the same topology, externally, these objects look significantly different. The simplest examples show the applicability of the developed approach in the design of technical products, in particular, with its help it is possible to draw a wing profile or a wheel with spokes. At the end of the work, screen shots of rather exotic figures are presented, which would have required considerable effort by other means. In the end, it is con­cluded about the possibilities of wide practical application of this approach.