This paper presents experimental and numerical investigations on the fire behaviour of GFRP-reinforced concrete (RC) slab strips. In the first part of this study fire resistance tests were performed on four concrete slab strips reinforced with sand-coated GFRP bars, in which the influence of the following parameters was assessed: (i) the concrete cover thickness; (ii) the presence of cold anchorages in the GFRP rebars, (iii) the presence of lap splices in the fire-exposed length and (iv) the concrete strength. The specimens were tested in a four-point bending configuration, being subjected to a sustained service load, while the bottom surface was exposed to the ISO 834 fire curve. These experiments were then complemented with the development of three dimensional (3D) thermo-mechanical finite element (FE) models of the slab strips to simulate the fire resistance tests. Both the temperature-dependent thermo-physical and mechanical properties of the constituent materials were considered; the GFRP-concrete interaction was also modelled by means of local bond vs. slip laws, which were previously calibrated by the authors for different temperatures. Comparisons between numerical and experimental results confirmed the accuracy of the FE models in predicting the thermo-mechanical response of GFRP-RC slab strips during fire exposure, namely in terms of temperatures, midspan deflection increase with time, failure modes and fire resistance. Both experimental and numerical results confirmed that (i) the fire resistance can be drastically reduced when the rebar splices are directly exposed to heat, and (ii) even adopting relatively small concrete cover thicknesses it is possible to attain considerable fire endurances provided that the anchors of the GFRP rebars remain sufficiently cold. Moreover, the specimen manufactured with a higher concrete strength presented less extensive cracking, reducing the localized heating of the reinforcement and leading to a higher fire resistance.