The purpose of this article is to present an Ultra-Wide Band GPR (ground-penetrating radar) technology in combination with an improved simplified Full-Waveform Inversion (FWI) approach for evaluation of bridge deck waterproofing membrane. gprMax model is used as a numerical forward model applied on various case studies corresponding to experimental configurations made in the laboratory. FWI has been designed and simplified for the 2D model. The structure under consideration is one-dimensional. This study have proved that the method is applicable to various multi-layered media configurations (various materials and structures). In this problematic, the FWI approach allows to evaluate and monitor the required dielectric and geometric parameters of the structure from a temporal signal (A-scan) with high accuracy and less time consuming. This punctual evaluation is needed for the next processing step, that is not discussed in this paper, based on Machine Learning method to detect automatically defects in bridge deck waterproofing membranes.