Wall paintings have a very long history, and during the Renaissance, paintings were mainly used in religious buildings. With time, the paintings beome increasingly vulnerable to physical and environmental degradation. Effective preservation requires detailed knowledge of the state of conservation of wall surfaces to guide operational decisions. Traditional GPR applications in this field are often hindered by the limited thickness of plaster and the complexity of near-surface features. This occurs because the pulse width at standard frequencies often exceeds the layer thickness, causing the surface reflection to mask the targets. In the Bardi Chapel (Basilica of Santa Croce), Giotto’s XIV-century murals depicting the life of St. Francis are currently undergoing a restoration campaign conducted by the Opificio delle Pietre Dure. During the restoration, a GPR framework was used to characterize the plaster layers that support these paintings. By employing a non-invasive NDT approach based on electromagnetic wave propagation, the research provides a high-resolution 2D and 3D imaging solution to detect internal defects without compromising the integrity of the historic material. Considering the limited thickness of the plaster and painted layers, the tests were conducted using the C-Thrue XS, a high-resolution GPR antenna with a 2 GHz central frequency with crossed polarization. This setup is designed to enhance the visibility of subtle internal anomalies while ensuring the physical integrity of the fragile surface. During data acquisition, a transparent sheet of paper was placed over the painting to prevent contact between the GPR antenna and the painted surface. This study aims to address these limitations, applying advanced GPR processing and interpretation strategies for improving the detection and characterization of surface defects and possible material heterogeneities in the wall. By optimizing the antenna's performance and data interpretation strategies, the framework provides a robust tool for the non-invasive monitoring of fragile murals.