Persons with serious mental illnesses experience high rates of medical comorbidity, especially diabetes. This study examined initial implementation feasibility, acceptability, and appropriateness of a new 6-month Multicomponent Intervention for Diabetes risk reduction in Adults with Serious mental illnesses (MIDAS) among persons in residential care facilities (RCFs). We conducted a mixed-methods study using four types of quantitative and qualitative data sources (administrative data; structured facility-level observations; resident assessments including blood-based biomarkers, 24-h dietary recalls, and self-report physical activity; and focus groups/interviews with staff and participants), to assess evidence of and factors affecting intervention feasibility, acceptability, and appropriateness. It was feasible to provide a high percentage of MIDAS class sessions (mean 50 of 52 intended sessions delivered) and make nutrition-related RCF changes (substitutions for healthier food items and reduced portion sizes). Class attendance rates and positive feedback from residents and staff provided evidence of MIDAS acceptability and appropriateness for addressing identified health needs. The residents who attended ≥ 85% of the sessions had greater improvement in several desired outcomes compared to others. Implementing a fully integrated MIDAS model with more extensive changes to facilities and more fundamental health changes among residents was more challenging. While the study found evidence to support feasibility, acceptability, and appropriateness of individual MIDAS components, some challenges for full implementation and success in obtaining immediate health benefits were also apparent. The study results highlight the need for improving health among RCF populations and will inform MIDAS adaptations designed to improve intervention fit and effectiveness outcomes.