<p>Engaging residents with the support available at community-based residential mental health rehabilitation facilities is an ongoing challenge for health services. This study explored factors associated with residential rehabilitation engagement across Queensland, Australia through regression modelling of cross-sectional data from a statewide benchmarking activity completed in 2023 (<i>n</i> = 208). The Residential Rehabilitation Engagement Scale (RRES) assessed each resident’s rehabilitation engagement. A broad range of potential predictors were considered, including resident and unit-level variables. Only 45.2% of residents had an average RRES score consistent with being engaged with rehabilitation support usually or always. Higher levels of rehabilitation engagement were significantly associated with lower levels of psychosocial disability (<i>B</i> = − 0.413, <i>p</i> &lt; .001), length of treatment (<i>B</i> = − 0.165, <i>p</i> = .008), care under the integrated staffing model (<i>B</i> = 0.156, <i>p</i> = .012), higher staff recovery knowledge and attitudes (<i>B</i> = 0.138, <i>p</i> = .037), and physical illness or disability (<i>B</i> = 0.129, <i>p</i> = .045). In conclusion, engagement in residential rehabilitation was associated with both resident and staff factors. The observation that engagement was higher where unit staff endorsed recovery knowledge and attitudes, and under the integrated staffing model is important. This suggests potential modifiable service characteristics that may support improved rehabilitation engagement in the future.</p>