Abstract: This study aimed to investigate the relationships between stress, sleep reactivity, dysfunctional beliefs about sleep and to estimate the distinct roles of occupational and family stress on sleep among military populations in China. A cross-sectional study was conducted with a sample of 389 members of the armed forces stationed on islands in China. The Likert scales, Pittsburgh Sleep Quality Index (PSQI), Ford Insomnia Response to Stress Test (FIRST) and Dysfunctional Beliefs and Attitudes about Sleep Scale (DBAS) were used to measure occupational and family stress, sleep quality, sleep reactivity, and dysfunctional beliefs about sleep, respectively. Mediation and moderation effects were tested using logistic regression. The indirect path between occupational stress and PSQI via FIRST scores was significant (ab = 0.564, p < 0.001). Mediation accounted for 38.2% of the total effects. The indirect path between family stress and PSQI via FIRST scores was also significant (ab = 0.295, p < 0.001). The mediating effects accounted for 100%. The mediation effects of sleep reactivity on the association between occupational or family stress and sleep quality became attenuated at one standard deviation (SD) above the mean and one SD below the mean of DBAS scores. Occupational stress can affect sleep quality both in direct and indirect paths, whereas family stress only affects sleep quality through the sleep reactivity system. Geographical factors may prevent individuals on islands from facing family stress directly. These results may open a new avenue for studying the mechanisms of sleep reactivity and ways to prevent sleep disturbance.