Several meta-analyses including a small number of cohorts showed inverse associations between the Mediterranean Diet (MedDiet) and risk of stroke. However, it remains unclear whether such a relation varies by region of the study population or by major subtypes of stroke. We searched PubMed and EMBASE databases for relevant studies and we further included unpublished results from the Singapore Chinese Health Study (N = 57,078) and the Seguimiento Universidad de Navarra (SUN) study (N = 12,670). We used a random-effects model to calculate summary relative risk (RR) with 95% confidence intervals (CI) of stroke for each 4-point increment of the MedDiet score, roughly corresponding to the difference between extreme quintiles of the MedDiet score among participants of the included studies. The final analyses included 20 prospective cohort studies involving 682,149 participants and 16,739 stroke cases. The summary RRs for each 4-point increment of the MedDiet score were 0.84 (95% CI 0.81-0.88; I2 = 11.5%) for all combined, 0.76 (95% CI 0.65-0.89) for studies in Mediterranean populations and 0.86 (95% CI 0.83-0.89) for those in non-Mediterranean populations. Lower risk of stroke associated with higher MedDiet score also was observed in the analyses stratified by study population and methodological characteristics including study risk of bias, version of the MedDiet index, and definition of moderate alcohol consumption. The MedDiet was similarly associated with lower risk of ischemic stroke (RR 0.86, 95% CI 0.81-0.91; nine studies) and hemorrhagic stroke (RR 0.83, 95% CI 0.74-0.93; eight studies). Our meta-analysis suggests that adhering to the Mediterranean diet was associated with lower risk of stroke in both Mediterranean and non-Mediterranean populations, and for both ischemic stroke and hemorrhagic stroke risk.