This paper presents the results of an experimental study aimed at evaluating the effect of important design parameters on the flexural behavior of unbonded post-tensioned rectangular concrete-filled fiber-reinforced-polymer (FRP) tubes (CFFTs) beams. Four full-scale beams were constructed and tested to failure in four-point bending over a simply supported span of 3,000 mm. The design parameters are the confinement effect of using FRP tube or steel stirrups, the concrete compressive strength, and the FRP tube thickness. The results showed that reinforced concrete (RC) rectangular CFFT beams prestressed with unbonded steel strands can achieve substantially higher flexural strength, inelastic flexural deformation, ductility, and energy absorption capacity than conventional prestressed RC beam without FRP tube. The ultimate strength, deformation capacity, ductility, and energy absorption were increased by 2.43, 1.89, 1.63 and 3.87 times than the prestressed RC beam. The results also showed that increasing the FRP tube thickness of the prestressed CFFT beams from 6.0 to 7.4 mm has led to the increase of the ultimate capacity, ductility, and energy absorption by 19%, 28% and 18%, respectively. Prestressed CFFT beams with high strength concrete (HSC, 70 MPa) enhanced the initial stiffness before cracking, increased the ultimate flexural moment capacity by 7%, with no significant change in the ultimate deflection, compared to normal strength concrete (NSC, 46 MPa). Based on the results of this study, the effect of increasing the FRP tube thickness is more effective in enhancing the flexural strength and stiffness of the prestressed CFFT beams than increasing the concrete strength. An analytical model based on plane sectional analysis using partially confined and unconfined concrete models as well as an empirical design equation are proposed to predict the flexural moment capacity of the tested beams. The proposed models using partially confined and unconfined concrete models and empirical equation successfully predict the flexural moment capacity with satisfactory accuracy on average of 1.06 ± 0.03, 1.23 ± 0.04 and 1.01 ± 0.08, respectively. It was also found that ignoring concrete confinement would highly underestimate the flexural strength. More investigations, however, are needed to assess the effect of a wide range of key influencing parameters to better model the flexural behavior of prestressed rectangular CFFT beams.