Background: Flexible assertive community treatment (FACT) is a community-based treatment model for patients with severe mental illness that has been widely implemented despite little evidence for its effectiveness. We aimed to evaluate the effect of FACT on mental health care outcomes compared with treatment from standard community mental health teams (CMHTs) or assertive community treatment (ACT) teams in Denmark. Methods: We did a quasi-experimental, propensity-score matched, controlled study in five FACT teams, four ACT teams, and five CMHTs in the Capital Region of Denmark. The FACT teams were established on May 1, 2016 (the index date). Patients were identified through the Danish Psychiatric Register and the study population consisted of all patients receiving care from any of these teams on the index date. Assignment to treatment was based on administrative considerations and was not done by the researchers. Patient data were collected from Danish registers. The primary outcome was analysed in the intention-to-treat population and compared mental health care outcomes (psychiatric admissions and bed days, outpatient contacts, self-harm, coercion, and death by any cause) for patients under the care of CMHTs or ACT teams that were reconfigured to FACT teams (CMHT–FACT or ACT–FACT) with patients who remained under the care of CMHTs and ACT teams. Patients who received FACT were matched using propensity scores with control patients from CMHTs and ACT teams to balance differences in baseline characteristics. Findings: On May 1, 2016, 2034 individuals (887 in the CMHT–FACT group, 887 in the matched CMHT group, 130 in the ACT–FACT group, 130 in the matched ACT group) were enrolled and were followed up from Nov 1, 2016, to Nov 1, 2018. The number of outpatient contacts was higher for patients receiving FACT than for those in the control groups (CMHT–FACT vs CMHT: incidence rate ratio 1·15; 95 % CI 1·10–1·20; ACT–FACT vs ACT: 1·15; 1·03–1·29). Patients receiving FACT had fewer admissions than those in the control groups (CMHT–FACT vs CMHT: 0·84; 0·76–0·92; ACT–FACT vs ACT: 0·71; 0·59–0·85). However, there were no significant differences in total inpatient days, use of coercion, episodes of self-harm, or deaths. Interpretation: To our knowledge, this is the first study to investigate the effect of FACT compared with treatment from a CMHT or ACT team. Our results suggest that FACT can provide a more intensive approach in terms of increased outpatient contacts than CMHT care or ACT. FACT requires further evaluation through randomised controlled trials that include a cost-effectiveness component before wider implementation. Funding: Mental Health Services in the Capital Region of Denmark.