The lack of inhibitory control toward foods may cause unhealthy eating behavior and lead to obesity. However, previous research failed to reach consensus on the alterations in event-related potential (ERP) markers of inhibitory control, i.e., N2 and P3. We hypothesized that the ERP effects of inhibitory control reported in previous food-based Go/NoGo studies might be obscured by non-inhibitory processes associated with stimulus probability. We designed two food-based Go/NoGo tasks composed of stimuli with the same type and frequency of occurrence (60% non-foods, 20% high-calorie foods, 20% low-calorie foods), one with response inhibition toward high- and the other toward low-calorie foods. Such an experimental design allowed us to isolate neural activity associated with inhibitory control from that associated with non-inhibitory processes by constructing ERP difference waves between NoGo and Go trials with the same frequency of occurrence. Electroencephalography data were collected from 32 obese participants and 29 normal-weight controls. Obese participants showed significantly lower accuracy in NoGo trials than normal-weight controls in both tasks. ERP data suggested inhibition-related effects for P3 (P3d) but not N2 in the difference waves, and obese participants showed significantly decreased P3d amplitudes than normal-weight controls in both tasks. In addition, we found that across both groups, individuals with larger waist-to-hip ratios showed smaller P3d amplitudes in both tasks, while such correlations between body mass index and P3d amplitude were only observed in the high-calorie task. Our findings suggest that the decreased effect of P3, not N2, might reflect the neural substrate of inhibitory control deficits in obese people. Thus, P3 could serve as an important neural marker in the future development of new therapeutic strategy that aims to improve inhibitory control in obesity.