To provide an up-to-date assessment of the effectiveness of the Mental Health First Aid (MHFA) training program on improving mental health knowledge, stigma and helping behaviour.Systematic review and meta-analysis.A systematic search of electronic databases was conducted in October 2017 to identify randomised controlled trials or controlled trials of the MHFA program. Eligible trials were in adults, used any comparison condition, and assessed one or more of the following outcomes: mental health first aid knowledge; recognition of mental disorders; treatment knowledge; stigma and social distance; confidence in or intentions to provide mental health first aid; provision of mental health first aid; mental health of trainees or recipients of mental health first aid. Risk of bias was assessed and effect sizes (Cohen's d) were pooled using a random effects model. Separate meta-analyses examined effects at post-training, up to 6 months post-training, and greater than 6 months post-training.A total of 18 trials (5936 participants) were included. Overall, effects were generally small-to-moderate post-training and up to 6 months later, with effects up to 12-months later unclear. MHFA training led to improved mental health first aid knowledge (ds 0.31-0.72), recognition of mental disorders (ds 0.22-0.52) and beliefs about effective treatments (ds 0.19-0.45). There were also small reductions in stigma (ds 0.08-0.14). Improvements were also observed in confidence in helping a person with a mental health problem (ds 0.21-0.58) and intentions to provide first aid (ds 0.26-0.75). There were small improvements in the amount of help provided to a person with a mental health problem at follow-up (d = 0.23) but changes in the quality of behaviours offered were unclear.This review supports the effectiveness of MHFA training in improving mental health literacy and appropriate support for those with mental health problems up to 6 months after training.PROSPERO (CRD42017060596).