In this study, the performance of secondary and tertiary carbonated water injection (CWI) was investigated at various operating conditions through sequences of laboratory experiments and numerical simulations. According to the results of CO2 solubility tests measured at constant temperatures, an increase in solubility value is observed for both water and oil phases when the pressure increases in the range of P = 0.7–10.3 MPa. Furthermore, it was found that the solubility of CO2 reduces with increased temperature. In addition, the results obtained from swelling/extraction tests revealed that the oil swelling factor increases as the pressure rises until a certain pressure called extraction pressure, Pext is reached. Afterward, the swelling factor reduces due to substantial extraction of lighter hydrocarbons from the oil to the CO2 phase. Comparison of the CO2 solubility values in oil at extraction pressures corresponding to different experimental temperatures also showed that the major hydrocarbon extraction occurs when a certain amount of CO2 has dissolved in the oil phase, which is called threshold CO2 solubility, χth. Results of sand-pack CWI flooding tests revealed that the recovery factor (RF) substantially increases up to a pressure of P = 5.6 MPa followed by slow growth until the pressure reaches P = 10.3 MPa. The same turning point of about P = 5.6 MPa was also observed in the plot of CO2 solubility in brine versus the operating pressure. Therefore, the value of CO2 solubility in brine controls the efficiency of the CWI. Additionally, lower recovery factor was obtained when temperature was increased from T = 25 °C to 40 °C. The same impact was observed when the carbonation level of the injected brine was reduced from CL = 100% to CL = 50%. From a CO2 storage point of view, the amount of CO2 stored by the end of secondary and tertiary CWI for different operating pressures was determined, and the values ranged from 40.7% to 61.1% of total injected CO2. Thus, it was concluded that CWI has great potential for permanent storage of the injected CO2 while significantly improving oil recovery in light oil systems. It was observed that the tuned PR-EOS model constructed using CMG WINPROP is capable of accurately reproducing the fluids’ basic characteristics, as well as the properties of CO2–oil and CO2–brine mixtures, such as saturation pressure and solubility. The fluid model was incorporated into the compositional and unconventional reservoir simulator, CMG GEM, in order to reproduce the CWI flooding tests conducted in this study. The simulation results showed that the CWI process can be simulated using the commercial software, CMG, by modifying the fluid model and history matching the laboratory flooding tests. Understanding the details of oil recovery mechanisms during CWI is of great importance, and precise observations of the fluid-fluid and fluid-solid interactions were carried out in this study through visual micro-model displacement tests. It was found that major wettability trapping along with minor snap-off and pore-doublet trapping mechanisms were attributed to relatively high residual oil saturation after primary water flooding. However, it was observed that injection of carbonated water is able to favourably adjust the wettability of the utilized glass micro-model toward the water-wetting condition. The observations of the CWI process in the micro-model demonstrated the main recovery mechanisms contributing to improvement of the oil recovery are oil swelling and viscosity reduction as a result of CO2 mass transfer from the brine to the in-place oil phase. It was also revealed that the trapped oil ganglia can be produced by continued injection of carbonated water for higher pore volumes.