Objective: To assess whether carbon monoxide (CO) poisoning increases the incidence of dementia. Methods: We searched the Cochrane Library, PubMed, and EMBASE from inception to 14 August 2022. Two authors independently selected studies, assessed the quality of included studies, and extracted data. Any disagreement was resolved by discussion with a third author. Only cohort study with an enough follow-up period was included for systematic reviews and meta-analysis. Results: Thirty-three full texts were initially searched, but only three studies met our inclusion criteria, and they were comprised of 134,563 participants who were initially free of dementia. The follow-up period ranged from 9 to 12 years. We found that CO poisoning increased the risk of dementia incidence (adjusted hazard ratio 2.61, 95% confidence interval 1.56 to 4.36, p=0.0003). Subgroup analysis showed that the increased dementia risk was significant in males but not in females, and the highest risk was in young age group, followed by in middle age group, but not in the old one. Conclusion Overall the evidence from prospective cohort studies supported a link between CO exposure and an increased dementia risk, although all the included studies were limited to Taiwanese population.