In this paper, the viability of the CO2 huff 'n' huff process as a primary means to enhance low-pressure light oil recovery under real reservoir conditions is investigated experimentally. Eight series of corefloods, a total of 35 runs of gas injection, were conducted in a composite core to evaluate the effect of major factors on the performance of the CO2 huff 'n' huff process. The factors evaluated included the gas injection rate, pressure depletion rate, maximum injection pressure and chasing gas (N-2), minimum termination pressure, and soaking time, etc. The experimental results showed that the first three cycles of a four-or five-cycle operation were the dominant contributors to the total oil production, resulting in a recovery factor of as high as 28.75% OOIP. Furthermore, two parameters used in this work to evaluate the process response, the recovery factor and the average oil production rate, demonstrated sensitivity to different operating parameters. In addition, the oil production of each cycle was observed to be mainly produced by the early production stage, during which the pressure was depleted from the highest level to a certain low level (e.g., 3 MPa in this study). Moreover, using N-2 as the chasing gas was found to have great potential regarding improving the CO2 efficiency while maintaining the performance of operations at a favorable level. This improvement was more significant in the second and third cycles than in the first cycle of operations. Besides, it was found that, for the first cycle of operations, an intermediate soaking time (e.g., 3 h) has a significant advantage over a short soaking time (1 h) or a long soaking time (10 h) in terms of average cycle oil production rate. However, for the second-and third-cycle injections, 1 h vs. 3 h of shut-in was considerably more beneficial. (c) 2016 Elsevier Ltd. All rights reserved.