Introduction Cognitive behavioural family intervention (CBFI) may be an effective brief psychosocial intervention for people diagnosed with severe mental illness (SMI) and their families. No systematic review has summarized the effectiveness of CBFI. Aim This review aimed to systematically examine the trial evidence of the effectiveness of CBFI versus treatment as usual (TAU) on improving the outcomes of people diagnosed with SMI and their families. Method Eligible randomized controlled trials were identified from nine databases. Three investigators independently took part in selection of articles, data extraction and risk assessment. Pooled treatment effects were computed using random-effects models. Results Four studies consisting of 524 participants were included. The risk of bias was low–unclear in most areas. The pooled CBFI effect on four service user outcomes including overall positive symptoms, delusions, overall negative symptoms and general psychopathology was significantly improved at post-treatment, compared with TAU, whereas effects on hallucinations and insight were equivocal. Discussion The findings reveal that CBFI is superior to TAU in treating positive and negative symptoms immediately following the intervention. Implications for Practice Mental health nurses may practise CBFI to enrich the psychiatric nursing service and promote nurse-led intervention. However, there is currently no substantial evidence that the intervention is effective over the longer term.