In general, it is expected that concrete structures using Glass Fibre Reinforced Plastic (GFRP) rebars as reinforcement could have improved durability compared to normal steel reinforcement because of the corrosion resistance of the rebar. However, there are some aspects of the behaviour of the GFRP bars under high temperature that must be explored. The aims of this work are to predict the fire rating of the GFRP rebars when embedded in concrete elements by creating a model and to validate the model by full-scale experiments. The first part of this work evaluates the effects of alkaline environments on the rebar itself, the bond strength at interface between the concrete and the rebar, and the strength of the GFRP rebars at a range of different temperatures (20-120°C). The three types of GFRP rods investigated in this work were subjected to alkaline solutions at 60°C for three different exposure times i. e. 30 days, 120 days and 240 days. Tensile and flexural tests were carried out for the physico-mechanical characterisation on the treated GFRP rebars specimens. As the immersion period and temperature increased, the strength of the rebars decreased. Data obtained from the first part of the work were used to predict long-term performance of the GFRP rebar in fire. The effects of higher temperatures with time on GFRP reinforced concrete members were also studied experimentally in this work. As a result equations were developed. These were validated with the help of the fire tests carried out in second phase of this work on two full-scale GFRP reinforced concrete beams. The first beam was reinforced with GFRP made from thermoset resin and in the second GFRP made from thermoplastic resin was used. Shear reinforcement for the first beam were GFRP stirrups and for the second beam steel stirrups were used. Degradation of flexural and shear capacities due to fire was evaluated using the modified design codes which is based on assessment of the reduction in the initial strengths of concrete and GFRP reinforcement, resulting from the high temperatures developed inside the beam. A comparison of the results for each beam is presented. Fire resistance (load bearing capacity) of GFRP RC beams complied with British Standard BS 478. These results are published for the first time in this work. The predicted failure time using the model compares well with the fire test results. The 3 result also indicated that the basic fire model needed adjustment mainly due to a difference in the assumed and observed failure modes. The importance of data necessary for a more accurate model has been identified as a programme for future work.