Abstract: Patients with non-affective psychosis often lead unhealthy lifestyles. We performed a systematic review and meta-analysis on non-pharmacological RCTs for improvement of diet and physical activity in non-affective psychosis patients, including first-episode psychosis. A variety of outcomes was analysed, including metabolic, psychopathology, cognitive, functional and quality of life outcomes. Fifty-nine studies were included. An improvement in anthropometric measurements (BMI, weight, waist circumference) was observed post-intervention, persisting after follow-up. Post-intervention benefit was found also for psychotic symptoms severity (also persisting after follow-up), many cognitive domains and physical and global functioning and quality of life. Conversely, no effect was observed in relation to most blood metabolites, blood pressure and non-psychotic psychopathology and spontaneous physical activity. Improvement was generally larger for interventions including exercise, especially moderate/vigorous aerobic exercise, but follow-up maintenance was greater for psychotherapy interventions. Sensitivity analyses limited to chronic stages of psychosis and low risk of bias studies produced comparable results. Further studies are needed to design optimized interventions in this vulnerable population.