Objective <p>To investigate the predictive value of acute brain lesions (ABLs) for the clinical outcomes and delayed encephalopathy after acute carbon monoxide poisoning (DEACMP).</p> Methods <p>This is a retrospective cohort study, patients with acute carbon monoxide (CO) poisoning were divided into the ABLs group and non-ABLs group. Outcome measures included clinical cure, favorable prognosis, and the incidence of DEACMP. Multivariate logistic regression analysis was used to identify independent predictors of clinical outcomes, and subgroup analyses were performed to clarify the impact of ABLs on clinical outcomes.</p> Results <p>Multivariate analysis showed that ABLs (clinical cure: adjusted odds ratio [aOR] 0.35, 95% confidence interval [CI] 0.17–0.75; DEACMP [DEACMP: aOR 2.54, 95% CI 1.08–6.00) and time from onset to first hyperbaric oxygen therapy (HBOT) &gt; 6&#xa0;h (clinical cure: aOR 0.38, 95% CI 0.18–0.82; DEACMP: aOR 3.50, 95% CI 1.58–7.74) were associated with a lower clinical cure rate and a higher incidence of DEACMP. A Glasgow Coma Scale (GCS) score &lt; 9 (aOR 7.50, 95% CI 3.33–16.89) was an independent predictor of DEACMP, but was not associated with clinical cure. Regarding ABLs in different locations, white matter lesions were significantly associated with lower clinical cure (adjusted OR = 0.23, 95% CI 0.10–0.55), lower favorable outcome (adjusted OR = 0.34, 95% CI 0.17–0.68), and higher DEACMP development (adjusted OR = 3.17, 95% CI 1.59–6.32). In contrast, basal ganglia lesions were only significantly correlated with lower clinical cure (adjusted OR = 0.45, 95% CI: 0.21–0.96), but not significant with favorable outcome or DEACMP occurrence.</p> Conclusion <p>ABLs and time from onset to first HBOT were significant predictors of clinical cure and DEACMP in patients. A GCS score &lt; 9 was significantly associated with DEACMP but showed no definitive correlation with clinical cure.</p>