Several reviews suggest that positive affect is associated with improved longevity, fewer physical symptoms, and biological indicators of good health. It is possible that positive affect could influence these outcomes by promoting healthful cognitions and behaviours. The present review identified conceptual pathways from positive affect to health cognitions and behaviour, and used random effects meta-analysis to quantify the impact of positive affect inductions (versus neutral affect conditions) on these outcomes. Literature searches located 54 independent tests that could be included in the review. Across all studies, the findings revealed no reliable effects on intentions (d+ = -.12, 95% CI = -.32 to .08, k = 15) or behaviour (d+ = .15, 95% CI = -.03 to .33, k = 23). There were four reliable effects involving specific cognitions and behaviours, but little clear evidence for generalised benefits or adverse effects of positive emotions on health-related cognitions or actions. Conclusions must be cautious given the paucity of tests available for analysis. The review offers suggestions about research designs that might profitably be deployed in future studies, and calls for additional tests of the impact of discrete positive emotions on health cognitions and behaviour.