Extensive research has been conducted on the use of tyre-derived products (e.g. rubber crumb and granule) to replace aggregates in producing concrete (i.e. rubber concrete). However, rubber concrete has so far been mainly limited to non-structural applications due to its well-known disadvantages including the relatively low stiffness and strength as well as early cracking as a result of lack of proper bonding between rubber and the paste matrix. The weaknesses of rubber concrete may be minimised in a hybrid column through lateral confinement by a fibre-reinforced polymer (FRP) tube and longitudinal reinforcement by steel or FRP. This paper presents an experimental study on FRP-confined rubber concrete (FCRC), which covers a large range of replacement ratio (0-75% by volume) of fine aggregates and three thicknesses of FRP. The test results confirmed the effectiveness of FRP confinement in improving the axial behaviour of rubber concrete, and clarified the effects of the two important parameters (i.e. replacement ratio of fine aggregates and FRP thickness). The test results also show that the behaviour of FCRC can be significantly different from that of FRP-confined natural aggregate concrete (NAC) with the same unconfined strength and confinement stiffness. By the inclusion of a simple coefficient to consider the effects of rubber aggregates, the existing models for FRP-confined NAC can be modified to provide reasonable prediction of the test results of FCRC.