In this paper we illustrate a set of balancing and modelling experiments, with the aim of reconstructing the kinematic evolution of the Umbria-Marche fold and thrust belt. The study area comprises the M. Coscerno anticline (South-Eastern Umbria), an eastward verging box fold, which is overthrust onto the Poggiodomo syncline. Three W-E trending, geological sections have been constructed across the M. Coscerno anticline. The availability of recent, detailed (1:10.000 scale) geological maps, edited by the Regione Umbria, supplied the surface geology data set, which provided the basis for the geological sections. The Northernmost section crosses the northern termination of the anticline, while the southernmost termination crosses the culmination of the same fold. The work consists of two parts: the first part is dedicated to the geometrical analysis of the structures along the sections; the second part is dedicated to the kinematic analysis, through a set of forward modelling experiments. The main aims of the first part are: the calculation of the detachment depth, and the evaluation of the shortening along the sections. Moving from North to South, we observe remarkable geometrical variations within the fold structure: i) the progressive thickening of the hinge zone; ii) the development of the forelimb of the anticline, involving increasing volumes of rocks, with a consequent increase in the fold wavelength; iii) the increase in the shortening value, up to a maximum of 4.8 km (i.e. 36% of the initial length of the section); iv) the deepening of the detachment level. The first three points can be related to the growth of the anticline, since the northernmost section can be viewed as an early stage of the anticline evolution, while the southernmost one represents the final geometry. The increasing depth of the detachment level is somewhat surprising, because the detachment level should coincide with a unique, mechanically suitable stratigraphic horizon. In the second part of the work, using forward modelling techniques, we checked the kinematic compatibility of the observed Umbria-Marche structures with the models of Fault Propagation Folding (FPF) and Fault Bend Folding, recently proposed for the Umbria-Marche structures. For these experiments, we used different modules of a software (GeoSec 4.1, produced by Cogniseis), specifically designed for the construction of balanced cross-sections, which implements several kinematic algorithms, including flexural and vertical/oblique slip folding. The method offers remarkable advantages; first of all the possibility of checking the sequence of the proposed deformations with a dynamic reconstruction, each stage of which keeps the balancing condition. In general, the experiments successfully model the observed geological structures, but also show some remarkable differences: i) in the Fold Propagation Fault algorithm implemented in the software the footwall rocks are not involved in the deformation resulting in an underestimation of the shortening; ii) the flexural slip algorithms do not allow a completely realistic simulation of the anticline geometry, enabling the thickening of the hinge zone; iii) the available kinematic models failed to simulate some details of the observed deformations, deriving from the disharmonic deformation of the Jurassic sequence with respect to the overlying Cretaceous-Oligocene sequence. We suggest a model of kinematic evolution for the Umbria-Marche structures, that is compatible with the observed geometry and with the proposed kinematic algorithms, and is based on the following characteristics: - the nucleation of concentric (detachment ?) folds, later accomplished by thrust deformation (break-thrust model); - the existence of multiple detachments; - the disharmony between the deformation of the Jurassic sequence and those of the overlying Cretaceous sequence.