Oil production from tight and shale formations accounted for more than half of total U.S. oil production in 2015 (EIA, 2016). Such amount is expected to grow significantly as the active development of low permeability reservoirs continues. Since the primary recovery is becoming less effective and large amounts of oil are locked in matrix, IOR techniques should be considered to face such challenge. Many recent studies have evaluated the potential of gas injection in shale plays. However, few studies discussed the feasibility of water huff-n-puff process. This study evaluated the potential of cyclic water injection (CWI) in shale plugs and compared it with the performance of cyclic gas injection (CGI). Eagle Ford core samples were used in this study and saturated with dead shale oil. X-ray diffraction (XRD) technique was employed to analyze the mineral composition of the sample. To reduce clay swelling, 5 wt.% of potassium chloride (KCl) solution was used as the injection fluid in CWI tests. Two groups of experiments were performed to examine the effects of soaking time and injection pressure (P in) on the recovery performance. For one cycle of huff-n-puff, after reaching the P in , the core plug was soaked with water for a certain time, then the surrounding pressure was released to enter the production period. Under the same operating conditions, CGI tests were conducted to compare the two IOR performances. The CWI experimental results show that soaking time affects the recovery factor (RF) of a single cycle within a certain range, which was similarly concluded from the CGI tests. Injection pressure can affect the recovery performance of CWI significantly. For example, after performing 12 cycles of water huff-n-puff, the RF for tests with P in of 1,000 psi and 5,000 psi were 14% and 21%, respectively. When applying 5,000 psi, fractures were created on the plug and the size enlarged with further cycles operated. Comparing the two processes of CWI and CGI, different recovery characters can be observed. CWI showed the recovery potential at the first four cycles, and then the incremental oil recovery decreased dramatically in the subsequent cycles. By contrast, CGI presented a more continuous and steady recovery performance with high incremental RF, and then gradually diminished after seven cycles of huff-n-puff processes.