Carbon dioxide (CO2) injection processes are among the most promising enhanced oil recovery techniques based on their great potential to improve oil production while utilizing geological storage of carbon dioxide to reduce greenhouse gas emissions. Among various CO2 injection modes, cyclic CO2 injection (CO2 huff-and-puff) scenarios have seen significant increase in interest for the purpose of enhanced oil recovery (EOR) in both non-fractured and fractured reservoirs. Several operating parameters, including operating pressure, solvent (CO2) injection time, soaking period, water saturation, etc., affect the performance of this process. However, the number of studies that consider these parameters is relatively limited. In this study, the performance of cyclic CO2 injection under various operating conditions for a light crude oil system is experimentally investigated. First, a comprehensive experimental study on the phase behaviour of the crude oil–CO2 system was conducted. Thereafter, a series of cyclic CO2 injection tests was designed and carried out in non-fractured and fractured porous media to determine the impact of various parameters on the recovery efficiency of this process. For the cyclic CO2 injection tests conducted at operating pressures ranging from immiscible to near-miscible conditions, it was found that the oil recovery increases considerably with operating pressures and reaches near maximum value at miscible condition. However, beyond this range, where the operating pressure exceeds the minimum miscibility pressure, the oil recovery factor was almost constant and further increase in operating pressure did not improve the oil recovery effectively. In addition, although it was seen that a longer soaking period and the presence of connate water saturation are positive parameters that enhance the recovery performance of immiscible cyclic CO2 injections, these parameters do not have noticeable influence in miscible injection scenarios. Furthermore, the results showed that longer CO2 injection time does not enhance the oil recovery. Additionally, it was observed that the cyclic CO2 injection process has a great capacity for CO2 storage, and it was found that the CO2 storage potential is more efficient if the cyclic injection process is implemented at pressures near the minimum miscibility pressure. The asphaltene precipitation inside the rock sample and its subsequent permeability reduction due to the CO2 injection were examined. The amount of the precipitated asphaltene in the porous media is considerably higher during miscible injection scenarios resulting in drastic reduction of the oil effective permeability. The compositional analysis of the remaining crude oil in the core also demonstrated that the mechanism of light component extraction by CO2 is much stronger during miscible cyclic CO2 injection compared to immiscible injection. The effect of fractures in the porous media on the oil recovery of cyclic CO2 injection was investigated, and the results showed that the presence of fracture significantly improves the oil recovery during the process. The impact of fracture was found to be more effective during immiscible cyclic CO2 injection. In addition, the examination of fracture orientation showed that horizontal fracturing remarkably enhances the oil production, while no noticeable increase in oil production was observed when the orientation of fracture was vertical. The numerical simulation of the process also revealed that the oil recovery of cyclic CO2 injection gives larger benefits from greater fracture width together with the presence of more fractures inside the system through enlarging the contact area between the CO2 and oil in-place.