Short sleep duration has been linked to higher levels of aggression. To synthetize all available research on this association, a systematic review and meta-analysis was performed. We included observational and experimental studies, using various measures of sleep duration and aggression. Eighty eligible papers were identified, describing 82 studies comprising a total number of 76.761 participants. Meta-analysis of results was possible for 60 studies. Pooled observational results on the association between sleep duration and aggression showed a correlation estimate of −0.16 (95%CI -0.19, −0.12; I2 = 83.9%) and an odds ratio estimate of 1.83 (95%CI 1.47, 2.28; I2 = 0.0%). For experimental studies, the pooled Standardized Mean Difference after manipulation of sleep duration was −0.37 (95%CI -0.80, 0.05; I2 = 89.05%) for controlled designs and −0.34 (95%CI -0.54, −0.14; I2 = 89.05%) for pre-post designs. Effect estimates were stronger for individuals with psychological vulnerabilities and younger persons. Exclusion of studies with low methodological quality strengthened the effect estimate in experimental but not in observational studies. To conclude, short sleep duration is associated with higher levels of aggression, with observational research strongly supporting the association and experimental studies providing mixed results. More well-designed prospective and experimental studies are needed to establish causality and optimize treatment, especially for individuals with psychological vulnerabilities.