Objective The aim of this systematic review and meta-analysis of randomised controlled trials (RCTs) includes determining the effects of core training alone or combined with conventional therapy on trunk function, balance, and gait in stroke patients; analysing these effects considering the stroke stage and the core training type; establishing the methodological quality of the studies published to date; and knowing the best dosage and type of exercise these aspects. Data source Cochrane Library, Medline, Web of Science, Scopus, and Science Direct were searched from inception to January 2022. Review method A review and meta-analysis of RCTs about core intervention effects on trunk function, balance, and gait in stroke patients was carried out following the Preferred Items for Reporting in Systematic Review and Meta-Analyses guidelines. The Cochrane Collaboration tool was used to assess the risk of bias and internal validity of the included studies. Results Twenty-nine studies were included (1030 stroke patients). The mean age of the participants was 58.46 ± 9.89 years, and the average time since the stroke incident was 308.64 ± 175.52 days. The meta-analysis results showed significant improvements in core interventions for trunk function ( P ≤ 0.008) and balance ( P < 0.00001) but not for gait performance ( P = 0.11 in chronic and P = 0.06 in pooled cases). In relation to the differential meta-analysis between training performed on stable and unstable surfaces, no significant differences were found in trunk function ( P = 0.06) or balance measures ( P = 0.05). Conclusion Core training improves trunk function and balance in acute and chronic patients, but no changes were found in gait performance.