OBJECTIVE: Research is limited on whether circadian rhythm and sleep architecture alterations during acute intracerebral hemorrhage (ICH) influence patient outcomes. This study aims to characterize these changes and explore their association with clinical prognosis, offering new insights for diagnosis and treatment. METHODS: We enrolled 100 acute hemorrhagic stroke patients who underwent continuous, contactless sleep monitoring via a smart mattress for 3-5 consecutive days. Prognosis was evaluated at discharge using the modified Rankin Scale (mRS), and patients were classified into favorable or unfavorable outcome groups. Circadian rhythm parameters (IS, IV, RA) and sleep metrics (eg, total sleep time, sleep latency, REM latency) during day and night were compared between groups. Multivariate logistic regression identified independent prognostic factors, and ROC analysis evaluated their predictive value. RESULTS: Group comparisons revealed statistically significant differences in RA and nighttime sleep latency between the favorable and unfavorable prognosis groups (P < 0.05). Binary logistic regression analysis identified nighttime sleep latency as an independent predictor of functional outcome (95% CI: 1.066 ~ 1.128, P < 0.05), which remained significant after adjusting for potential confounders (95% CI: 1.016 ~ 1.148, P < 0.05). The mean nighttime sleep latency was 18.14 minutes in the favorable group and 12.30 minutes in the unfavorable group. The area under the ROC curve (AUC) for nighttime sleep latency was 0.642 (95% CI: 0.526-0.757, P = 0.028), with an optimal cutoff value of 10.95 minutes, yielding a sensitivity of 72.2% and specificity of 53.6%. CONCLUSION: Hemorrhagic stroke patients show disrupted circadian stability, with greater RA reductions in those with worse outcomes. Nighttime sleep latency independently predicts poor prognosis with moderate accuracy. Circadian rhythm stability may serve as a prognostic marker in hemorrhagic stroke to avoid implying causality.