Genetic variants of paraoxonase 1 (PON1) were implicated in stroke susceptibility in several case-control association studies. However, the studies have reported apparently conflicting results, rendering precise assessment of the disease risk associated with the variants difficult. A meta-analysis was therefore conducted by including the studies that examined the association between two common polymorphisms (L55M and Q192R) in the coding region of PON1 gene and the risk of stroke. Altogether 10 studies on L55M polymorphism and 11 studies on Q192R polymorphism were included in this meta-analysis. The results showed, although there was no significant association of the 55L allele with stroke [random effects OR = 1.09, 95% CI (0.93, 1.27), P = 0.29], the 192R allele conferred significant risk of stroke in the overall study population [random effects OR = 1.25, 95% CI (1.07, 1.46), P = 0.006]. Same pattern of results as both the allele contrasts was obtained for the homozygote contrasts and the dominant, recessive and additive models. Subgroup analyses for stroke type, age of patients and ethnicity revealed no association of the 55L allele with stroke, whereas the association of the 192R allele persisted significantly in the groups comprising ischemic stroke patients, stroke patients with mean age >60 years and Caucasian subjects. But no significant association of this allele with stroke susceptibility was detected in the East Asian population. Therefore, the results of this meta-analysis indicate, the Q192R polymorphism could be an important risk factor for stroke, especially in the Caucasian population.