Background People without stable housing—and Veterans specifically—are at increased risk of suicide. This study assessed whether unstably housed Veterans’ participation in homeless services is associated with reduced risk of all-cause and suicide mortality. Methods This retrospective cohort study used a sample of 169 221 Veterans across the US who self-reported housing instability between 1 October 2012 and 30 September 2016. Multivariable Cox regression models assessed the association between Veterans’ utilisation of homeless services and all-cause and suicide mortality, adjusting for sociodemographics and severity of medical comorbidities. Results More than one-half of unstably housed Veterans accessed homeless services during the observation period; utilisation of any homeless services was associated with a 6% reduction in hazards for all-cause mortality (adjusted HR[aHR]=0.94, 95% CI[CI]=0.90–0.98). An increasing number of homeless services used was associated with significantly reduced hazards of both all-cause (aHR=0.93, 95% CI=0.91–0.95) and suicide mortality (aHR=0.81, 95% CI=0.73–0.89). Conclusions The use of homeless services among Veterans reporting housing instability was significantly associated with reduced hazards of all-cause and suicide mortality. Addressing suicide prevention and homelessness together—and ensuring ‘upstream’ interventions—within the context of the VHA healthcare system holds promise for preventing suicide deaths among Veterans. Mental health treatment is critical for suicide prevention, but future research should investigate if social service programmes, by addressing unmet human needs, may also reduce suicide.