Timber-timber composite (TTC) floors have received considerable attention in recent years, thanks to environmentally friendly attributes and the high strength over weight ratio of the timber. The massive TTC floors consist of flat cross laminated timber (CLT) slabs connected to laminated veneer lumber (LVL) or glued laminated timber (GLT) beams. Despite numerous advantages, the mechanical behaviour of timber, and consequently stiffness and deflections of the TTC joints and beams are highly influenced by the time-dependent effects and environmental conditions, i.e., relative humidity (RH) and temperature. Accordingly, this project aimed at investigating the short- and long-term performance of the TTC systems under variable RH and temperature. Short-term pushout tests were performed on LVL/GLT beam to CLT slab joints (with coach screw shear connectors) to characterise the load-slip, peak load, and failure mode of the joints. Furthermore, material tests were conducted on timber and the screws, and the results were used to build a versatile beam on inelastic foundation finite element (FE) model and predict the full range load-slip response of the TTC joints. Comparison of the experimental peak loads with the FE predictions demonstrated the major influence of the rope effect on the load-slip and peak load of the TTC joints with screw shear connectors. To assess the long-term performance of the TTC system, twelve TTC joints and eight full-scale TTC beams were fabricated by connecting LVL/GLT beams to CLT slabs. The TTC joints and beams were subjected to sustained (constant) loads and monitored under an indoor uncontrolled environment for over 500 days. The test results revealed the major influence of the CLT slab thickness and orientation (crosswise or lengthwise), shear connector types/sizes and CLT slab continuity (segmentation) on the long-term performance of the TTC joints and beams. The creep coefficient of the TTC joints after 523 days was in the range of 1.15-4.25 and the creep coefficient of TTC beams after 540 days was in the range of 0.39-0.73. A creep model (calibrated against the experimental slip-time curves) was used to predict the creep coefficient for the 50-year design life of the TTC systems. Simplified analytical models were modified and utilised to predict the mid-span deflections of the TTC beams. The mid-span deflections predicted by the simplified models had good agreement with the test data. The inelastic strains due to (differential) shrinkage between the slabs and joists had a minor (less than 3%) influence on the long-term deflections predicted by the models. Lastly, a 3D fully coupled finite element (FE) model implemented as user-defined subroutines in ABAQUS/Standard was validated against the TTC joints and beams long-term test results and accordingly input parameters for long-term FE simulations of the LVL-CLT and GLT-CLT composites (with screw shear connectors) are recommended.