In continuous CO2 secondary flooding, severe viscous fingering and early CO2 breakthrough (BT) occur due to an unfavourable mobility contrast between the injected CO2 and the crude oil. The oil recovery is limited after CO2 BT, which is attributed to the gas channeling problem. In continuous CO2 tertiary flooding, the previously injected water helps to control the mobility of the subsequently injected CO2. On the other hand, it hinders the mutual interactions between the residual oil and the subsequently injected CO2 and thus affects the tertiary oil recovery. This thesis experimentally studied the mutual interactions between the light oil/reservoir brine and CO2 and the technical optimization of CO2 enhanced oil recovery (CO2-EOR) process, which included the optimum timing for CO2-EOR after waterflooding and soaking effect on miscible CO2 flooding in a tight sandstone formation. First, the saturation pressure (Psat), oil-swelling factor (SF), gas–oil ratio (GOR) of CO2- saturated light oil, and gas–water ratio (GWR) of CO2-saturated reservoir brine were measured by using a PVT system. Second, the viscosities of CO2-saturated light oil with different CO2 concentrations were measured by using a capillary viscometer. Third, the equilibrium interfacial tensions (IFTs) between the light oil/reservoir brine and CO2 were measured at different equilibrium pressures and the actual reservoir temperature by applying the axisymmetric drop shape analysis (ADSA) technique for the pendant oil drop case. Finally, five coreflood tests were performed to determine the optimum timing for miscible CO2-EOR after waterflooding and six coreflood tests were conducted to examine CO2-soaking effect on miscible CO2 flooding in a tight sandstone formation, respectively. The experimental results showed that Psat and oil SF of CO2-saturated light oil were increased respectively in the ranges of 4.97–8.44 MPa and 1.14–1.34 when CO2 concentration in the light oil–CO2 system was increased in the range of 38.94–60.46 mol.%. The measured GOR of CO2-saturated light oil was found to be approximately six times of the measured GWR of CO2-saturated reservoir brine at the same pressure and temperature. The respective viscosities of CO2-saturated light oils with 38.94 and 60.46 mol.% CO2 concentrations were reduced to lower than 16% and 10% of the original dead light oil viscosity at the same reservoir temperature. The measured equilibrium IFT between the light oil and CO2 was about one third to one tenth of that between the reservoir brine and CO2 under the same test conditions. By comparing the total oil recovery factor (RF) of waterflooding and CO2 flooding in terms of the original-oil-inplace (OOIP), the oil RF of CO2 flooding in terms of the residual-oil-in-place (ROIP), and the pore volume (PV) of CO2 BT from the beginning of CO2 flooding, it is found that the optimum timing for starting miscible CO2 tertiary flooding is when waterflooding reaches half of its maximum secondary oil RF. It was found that in the coreflood tests without CO2 soaking, a large amount of oil was produced in the first PV of CO2 injection, whereas only a low oil RF of 0.77–6.10% was obtained in the second PV of CO2 injection. In the coreflood tests with CO2 soaking, the composite sandstone reservoir core plugs with the residual light oil, reservoir brine, and remaining CO2 were soaked for 24 h after the first PV of CO2 injection. A high oil RF of 11.05–14.74% was achieved in the second PV of CO2 injection. Among the six CO2 coreflood tests, pre-waterflooding plus CO2 tertiary flooding with CO2 soaking resulted in the highest oil RF, which was attributed to the mobility-control effect of pre-waterflooding and CO2-soaking effect.