The potential of the CO2 soaking procedure has been generally acknowledged as a valid way to advance the tight rock oil recovery. Over the last decade, a significant number of Huff-n-Puff (H-n-P) experiments have been conducted to develop unconventional oil reservoirs. However, the majority of experiments used fully saturated cores and unconfined core holders. Therefore, the average oil recovery at the field-scale could not be accurately estimated. Besides, the effect of key factors such as fracture geometry on bypassed oil recovery has remained obscure. For better quantifying CO2 H-n-P efficiency in oil fields, this study proposes an immiscible CO2 soaking process aimed at bypassing the oil before conducting the H-n-P process using various fracture forms and dimensions. Tight cores from Sarawak with an average porosity and permeability of 9% and 0.07 md, respectively, were employed in this experimental research. The importance of the fracture surface areas (FSA), fracture depth (FD), width (FW), and diameter was thoroughly studied. The research findings revealed that the two-step CO2 soaking procedure significantly reduces the effectiveness of the currently applied laboratory H-n-P process. However, the outcomes are more consistent with the current average oil recoveries in field pilots. The study demonstrates that FD is the most critical factor in maximizing the remaining oil recovery. The research indicates that the FSA does not always follow a specific trend. It is, however, dependent on the fracture geometry. The significance of the crack's surface area and fracture intensity is determined to be primarily dependent on the fracture shape and the utilized core holder system, respectively. The study's findings presented a higher degree of accuracy in estimating actual oil recovery from tight reservoirs employing two-step soaking technology.