This study had two main purposes: (1) whether gratitude could predict the sense of coherence (SOC) via positive affect and positive coping style; (2) whether gender could moderate the mediating effect of positive affect and positive coping style. We hypothesized that gratitude could directly predict SOC (H1); gratitude could also enhance SOC through increasing positive affect (H2) and positive coping style (H3) ; Furthermore, positive affect and positive coping style played a chain mediating effect between gratitude and SOC (H4) ; gender could moderate the mediation model (H5). 449 participants were recruited from two ordinary high schools in Hohhot, 51.6% of which were male and 48.1% female (Mage= 17.58, SD= 1.06). The questionnaires were Adolescent Gratitude Scale, Adolescents’ School Positive and Negative Affect Scale, Trait Coping Style Questionnaire and SOC Scale (13 items). Furthermore, a confirmatory analysis was used to test common method variance. Results indicated that four-factor model was better than single-factor model(Δχ2/Δdf = 72.45, p < .001), making sure the study is accurate enough in terms of common method variance. In order to validate our hypotheses, a structural equation model (SEM) was built. Overall, the results of goodness-of-fit indices revealed that the SEM fitted the data well (χ2/df = 3.21, p < .001，RMSEA = .07, CFI = .94, NNFI = .91, GFI = .95). The Bias-corrected Nonparametric Percentile Bootstrap test showed that both the direct and indirect effect were significant. Specifically, gratitude could directly predict SOC (the effect size was 46%),?and more importantly, it could also enhance SOC through increasing positive affect (the effect size was 24%) and positive coping style (the effect size was 10%). Also, positive affect and positive coping style indeed played a chain mediating effect between gratitude and SOC (the effect size was 20%). Therefore, our hypotheses 1- 4 were verified. A multi-group analysis found that the CRD (Critical Ratios for Differences between Parameters) of path coefficient from gratitude to positive coping style was only significantly different between males and females (CRD = |-2.06| > 1.96, p < .05), so our hypothesis 5 failed to be fully validated. We concluded from this study that gratitude could not only predict SOC directly, but also indirectly influence SOC via positive affect and positive coping style, and males with high gratitude scores were more likely to use positive coping style than females. The main innovative point of this study had three aspects: (1) exploring the ways to enhance SOC, which filled a gap in the field of domestic SOC study, (2) helping to explain the relationship between gratitude and mental health, and (3) providing new evidence for the influence of gender on the function of gratitude. In addition, our findings also had practical implications for both gratitude intervention and gender role education. At last, we discussed the unaddressed field and pointed out the possible direction for future research.