Perceived stress has been associated with adverse health outcomes. Although people experiencing homelessness often report multiple acute and chronic stressors, research on resilience and perceived stress on the general homeless population is limited. This longitudinal study examined homeless adults with mental illness who were part of a 24‐month trial of Housing First to explore: (a) changes in levels of resilience and perceived stress during the trial, and (b) the association between levels of resilience and perceived stress with measures of social support, social functioning and percentage of days stably housed over the study period. This longitudinal study (2009–2013) that used trial data included 575 participants in Toronto, Ontario. Of these individuals, 507 were included in this study. Connor‐Davidson Resilience Scale and Perceived Stress Scales (PSS) measured the two outcomes, resilience and perceived stress. Time (baseline, 12 and 24 months), housing stability and three measures of social support and social functioning were the main predictors. A longitudinal analysis was done with repeated measures analysis of resilience and perceived stress using linear mixed models with random intercepts. Mean resilience scores increased (baseline: 5.1 [95% CI: 4.9, 5.2], 12 months: 5.5 [95% CI: 5.3, 5.7], 24 months: 5.6 [95% CI: 5.4, 5.8]), and PSS scores decreased (baseline: 22.3 [95% CI: 21.5, 23.0], 24 months: 18.6 [95% CI: 17.9, 19.4]). In the multivariable analyses, increased resilience was associated with higher scores on the three social support and social functioning measures, (estimates = 0.12, 0.04, 0.02) but not percentage days stably housed. Lower PSS scores were associated with higher scores on all three social support and social functioning measures (−0.20, −0.33, −0.21) and higher percentages of days stably housed (−0.015). Strong social support and social functioning may minimise the harmful effects of stressful life events on homeless individuals by increasing resilience and reducing stress. Interventions to help homeless people build appropriate support networks should be delivered in parallel to efforts that increase housing stability.