Although dietary restraint has been shown to be a robust predictor of binge eating among women, many women report elevated levels of dietary restraint but do not concurrently exhibit symptoms of binge eating. Moderating variables could therefore interact with dietary restraint to affect its relation to binge eating. One potential factor that may attenuate this relationship is eating-related self-efficacy, defined as the tendency to feel confident in the ability to control eating behaviour under a diverse set of circumstances (e.g., under negative affect, social conflicts). This cross-sectional study examined whether eating-related self-efficacy moderated the relationship between flexible (i.e., a graded approach to dieting, defined by behaviour such as taking smaller servings to regulate body weight, yet still enjoying a variety of foods) and rigid restraint (i.e., an all-or-none approach to eating, characterised by inflexible diet rules) and binge eating. Data were analysed from 237 women. Greater levels of rigid restraint, flexible restraint, and a poorer self-efficacy were shown to predict unique variance in binge eating severity. A significant interaction effect was observed between flexible (but not rigid) restraint and self-efficacy scores on binge eating. Contrary to expectations, however, the flexible restraint-binge eating relationship was largest for those with moderate to strong self-efficacy, and was non-significant for those with poor self-efficacy. Overall, findings suggest that different mechanisms may be operating to maintain binge eating in those with varying levels of eating-related self-efficacy.