Aim To determine the association between dementia and age-related macular degeneration (AMD) using meta-analysis. Methods We searched in the MEDLINE, EMBASE, Web of Knowledge, PsycInfo and Cochrane database of systematic reviews for studies published from March 1959 to March 2018. We included cross-sectional, case-control and cohort studies that evaluated the association of dementia/Alzheimer's disease (AD) with AMD (as outcome) and the association of AMD with dementia/AD (as outcome). Studies that compared cognitive functions between AMD and controls were also included. The summary outcomes, namely odds ratio (OR), relative risk, mean differences and corresponding 95% CIs, were estimated using random effects models. We performed sensitivity analysis based on study quality and individual study effect to control for potential biases. Results Among 2159 citation records, we identified 21 studies consisting of 7 876 499 study subjects for meta-analysis. Patients with dementia (p(adjusted)<= 0.017, OR >= 1.24, I-2 <= 9%) or AD (p=0.001, ORunadjusted=2.22, I-2=50%) were at risk for AMD, particularly for late AMD (p(adjusted)<0.001, OR=1.37, I-2=0). AMD was also significantly associated with increased risk of AD/cognitive impairment (p(adjusted)=0.037, OR=2.42, I-2=38%). Moreover, patients with AMD had poorer cognitive functions when compared with controls, including Mini-Mental State Examination (p<0.001, I-2 <= 79%) and Trail Making Test A (p<0.001, I-2=0). Sensitivity analysis and Egger's test indicated our results were less likely biased. Conclusions A significant association between dementia/AD and AMD calls for greater clinical awareness. The cost-effectiveness of routine screening for the other condition in patients with primary diagnosis of dementia/AD or AMD requires further study.