Gas Huff-n-Puff (HnP) has been implemented as an enhanced oil recovery technique to recover the residual oil after primary production from unconventional wells. Natural gas, nitrogen, and carbon dioxide can be used as candidates for injecting fluids in HnP operations. However, natural gas is the common choice for injecting gas due to its availability and incentives for reducing the venting and flaring of the produced gas. So far, some experimental studies attempted to investigate the natural gas HnP on Eagle Ford shale. However, the associated oil-recovery mechanisms are poorly understood. In this study, we perform natural gas HnP experiments using C1 and C1-C2 on Eagle Ford shale samples under representative reservoir conditions. We use a custom-designed visualization cell to observe the interactions of gas, oil, and shale during the whole HnP cycle. Consistent with field operations, we adopt a hybrid depletion strategy of steep depletion at the initial stages followed by a slow depletion at later stages. We select the pressure depletion rates by downscaling field data of a HnP operation in the Eagle Ford Formation. Our results reveal that solution-gas drive or gas expansion during the depletion stage is the dominant oil-recovery mechanism. However, the extended soaking period helps in oil recovery by allowing sufficient gas diffusion into the core plug. We observe that enrichment of injecting gas by C2 results in earlier and more oil production compared to pure C1. The ultimate oil recovery factor after a single-cycle C1 and C1-C2 HnP is 46.1 and 55.6% of the original oil-in-place, respectively. We estimate the apparent diffusivity coefficient of C1 and C1-C2 in oil-saturated shale plugs using available analytical models. The estimated apparent diffusivity coefficients are in the order of 10–10 m2/s with an 8% higher diffusivity coefficient in the case of C1-C2 compared to the case of C1.