Intimate partner violence (IPV) is a leading cause of homelessness and a serious threat to public health and well-being. To determine whether the Domestic Violence Housing First (DVHF) model improves safety, housing stability, and mental health over 2 years. This longitudinal comparative effectiveness study interviewed IPV survivors and reviewed their agency records. All unstably housed or homeless IPV survivors entering domestic violence (DV) services were eligible to participate in the study, ensuring capture of typical variability in service delivery (eg, some survivors would enter services when agencies had the capacity to provide DVHF and others would receive services as usual [SAU]). Clients from 5 DV agencies (3 rural and 2 urban) referred by agency staff in a Pacific Northwest state of the United States were assessed between July 17, 2017, and July 16, 2021. Interviews were conducted in English or Spanish at entry into services (baseline) and at 6-, 12-, 18-, and 24-month follow-up visits. The DVHF model was compared with SAU. The baseline sample included 406 survivors (92.7% of 438 eligible). Of the 375 participants retained at the 6-month follow-up (92.4% retention), 344 had received services and had complete data across all outcomes. Three hundred sixty-three participants (89.4%) were retained at the 24-month follow-up. The DVHF model has 2 components: housing-inclusive advocacy and flexible funding. Main outcomes included housing stability, safety, and mental health, which were assessed using standardized measures. Of the 344 participants (mean [SD] age, 34.6 [9.0] years) included in the analyses, 219 (63.7%) received DVHF and 125 (36.3%) received SAU. Most participants identified as female (334 [97.1%]) and heterosexual (299 [86.9%]). Two hundred twenty-one participants (64.2%) were from a racial and ethnic minority group. Longitudinal linear mixed-effects models showed that receiving SAU was associated with greater housing instability (mean difference, 0.78 [95% CI, 0.42-1.14]), DV exposure (mean difference, 0.15 [95% CI, 0.05-0.26]), depression (mean difference, 1.35 [95% CI, 0.27-2.43]), anxiety (mean difference, 1.15 [95% CI, 0.11-2.19]), and posttraumatic stress disorder (mean difference, 0.54 [95% CI, 0.04-1.04]) compared with receiving the DVHF model. Evidence in this comparative effectiveness study suggests that the DVHF model was more effective than SAU in improving the housing stability, safety, and mental health of survivors of IPV. The DVHF's amelioration of all of these interconnected public health issues-relatively quickly and with long-term continuance-will be of substantial interest to DV agencies and others working to support unstably housed IPV survivors.