The psychological benefit of gratitude has been well demonstrated in previous studies. However, when we examined these studies closely, we found that the moderators were rarely investigated, suggesting that further work is needed to explore the boundaries of gratitude In this regard, the authors have proposed that ambivalence over emotional expression might be a potential moderator that would inhibit the beneficial effect of gratitude on well‐being. Two studies were conducted to examine our hypothesis. Study 1 consisted of 353 Taiwanese college students who completed the Gratitude Questionnaire–Taiwan version (GQ‐T), Ambivalence over Emotional Expression Questionnaire (AEQ), and one question about subjective happiness. We found that ambivalence over emotional expression significantly moderated the effect of gratitude on happiness. To validate our findings in Study 1, 233 Taiwanese college students were recruited for Study 2, and they completed the GQ‐T, AEQ, subjective happiness short‐form UCLA loneliness scale, as well as the Center for Epidemiological Studies Depression Scale (CES‐D). Both studies demonstrated that ambivalence over emotional expression moderated the relationship between gratitude and well‐being indexes. Simply stated, the authors found that across the two independent samples, among students who are high in ambivalence over emotional expression, the beneficial effect of gratitude on subjective happiness was inhibited. However, the moderating pattern for loneliness and depression was contrary to our expectations, indicating that high ambivalence over emotional expression does not inhibit gratitude. Possible explanations and implications for social relationships and emotional expression are discussed.