Previous studies have examined the role of three NOS3 gene polymorphisms [G894T, T-786C, and the variable number of tandem repeats 4b/a (VNTR 4b/a)] in the susceptibility to preeclampsia with inconclusive findings. We therefore conducted an updated meta-analysis by including more studies. The most appropriate genetic model was chosen for each polymorphism by using a well-established method. Pooled results indicated that, compared with the GT + GG genotype, the TT genotype of G894T was associated with an increased risk of preeclampsia (odds ratio (OR) = 1.46; 95% confidence interval (CI) = 1.21–1.77, P < 0.001; I2 = 40.2%). The CC genotype of T-786C was also associated with a higher risk of preeclampsia (OR = 1.30; 95% CI = 1.07–1.58, P = 0.034; I2 = 46.9%) than the CT + TT genotype. No association was found for VNTR 4b/a. Stratified analysis indicated that the increased risk was evident for high-quality studies both for G894T and T-786C, and for studies conducted among Caucasians and Africans for T-786C. However, the increased risk for T-786C among Africans needs further confirmation due to the high probability of false-positive reports. Our results suggested that G894T and T-786C polymorphisms, but not VNTR 4b/a, were associated with an increased risk of preeclampsia.