The purpose of this meta-analysis was to examine the current effectiveness of physical activity (PA) interventions to change affective judgements (AJ) and subsequent behaviour and explore potential moderators. Eligible studies were published in a peer-reviewed English journal and included an experimental design in the PA domain with a measure of AJ as the dependent variable, among adults (>17 years). Literature searches concluded in July 2017 using 11 common databases, with additional hand searching conducted in February 2018. The search yielded 32 independent studies. Random-effects meta-analysis showed positive changes in AJ favouring intervention over control groups, g = 0.43 (95% CI = 0.26-0.60). These changes predicted (β = 0.64) positive changes in PA, g = 0.38 (95% CI = 0.16-0.60), among a sub-sample (k = 14) of studies that also provided behavioural data. Moderator analyses showed the effects were inflated by potential publication bias, participant gender, baseline PA and focus of the intervention. AJ may show change from intervention but larger sample studies are required to obtain a more reliable effect size estimate. Further, few studies have employed behaviour change techniques that would align with the theoretical reasons for changes in AJ, so our evidence for practical intervention content is limited. [ABSTRACT FROM AUTHOR]