Substance addiction and food addiction are significant social problems worldwide. In previous studies of substance addiction, transcranial direct current stimulation (tDCS) has been used to influence craving of substance or food. However, the reported effects are not always consistent due to inconsistent experimental settings. The way modulators influence the effect of tDCS on substance addiction is worth exploring. This meta-analysis was conducted to estimate the effect size of tDCS on substance and food craving and to investigate the influence of potential modulators. We systemically identified and reviewed studies on substance/food craving using tDCS that were published between January 2008 to January 2020. A total of 32 eligible studies were identified. Hedges' g was computed as an indicator of the effect of tDCS and some potential moderators (substance type, stimulation sites, current intensities, number of sessions, duration of stimulation, and study design) were examined using subgroup analysis. Random effects analysis revealed a total medium effect size [Hedges'
g
= 0.536, 95% confidence interval (CI): 0.389–0.683, after adjusting Hedges'
g
= 0.416, 95% CI: 0.262–0.570] preferring active over sham stimulation to reduce craving. A significant difference was observed between the number of sessions (repeated stimulation was better than single stimulation). The duration of stimulation may have a positive influence on the effects of tDCS. No other significant differences were found in other subgroups analysis. In conclusion, our results provided evidence that tDCS can be an effective way to reduce craving of substance or food, and longer multiple stimulus durations in all can more effectively reduce craving; however, the influences of modulators still need be to be examined in depth in future.