The task of creating a 3D model of a real object is topical in many spheres of science and technology. Different mathematical descriptions exist in computer graphics, which allow describing the free-form surface of a real object. One of the most efficient approaches is the description of analytical surfaces in the form of parametric surfaces. The task of real object geometrical 3D model construction in analytical form, based on the array of 2D input data, can be divided into two parts: 1) selection of data (control points) in 2D slice and creation of an array of input data for 3D modelling; 2) the construction of the real object’s 3D model in the analytical form. In this work, the authors propose a modification of control point parameter calculation in the method of control point selection in 2D slices, and a new method of 3D model construction of a real object using interpolation curves in two parametrical directions. The modification of the calculation of the control point parameter is connected with the deformation of the adaptive contour in such a way that the line from the coordinate’s centre intersects the contour more than once. In this case, additional curve points of the adaptive contour are found and the parameter values are distributed evenly between these points. Then, using the parameter values and control point coordinates an interpolating parametrical curve is constructed. This is the first step of the 3D model construction. The initial data for the second step of 3D model construction is an array of parametrical curves; each curve interpolates an adaptive contour in 2D slice. Using this data by analogy with points on the adaptive contour an interpolating parametrical surface of the researched object. In this work an approbation of the proposed method is conducted on the real medical object and the method is compared with the existing method [5]. The comparative analysis shows that the proposed approach gives better results from the point of model’s quality.