Institutional violence presents significant challenges to the accomplishment of legitimate social order in prison. This systematic review examines the effect of psychoeducational programs on violent behaviour in prison. Comprehensive searches of the empirical research literature were conducted to identify randomized and non-randomized studies carried out in the last two decades (1996–2016) that compared psychoeducational programs with treatment as usual (TAU). The content of programs was analysed and classified. The design of the studies was subject to a risk of bias analysis and quality assessment. Violent behaviour in prison was measured by institutional reports, inmate self-reports, observer ratings, or using psychometrically-valid scales. We identified 21 separate studies with considerable variations in program quality and evaluation methodology. The majority of programs adopted a cognitive behavioural or social learning approach. There was limited evidence for the efficacy of these programs, although highly-structured programs showed the most promise. Programs that aimed to integrate their treatment ethos into the institutional regime and target specific criminogenic risks also produced evidence of effectiveness in reducing institutional violence. The current evidence base does not provide a clear answer to the ‘what works’ question in reducing institutional violence. However, there is evidence that some approaches are more successful than others and this should guide future program design and evaluation.