Abstract： Objective To evaluate the relationship between thyroid autoimmunity and the risk of preterm birth. Methods Literature search was done among PubMed, Embase, Wanfang Medical Database, China Academic Journal Network Publishing Database and China Biology Medicine disc from Jan. 1st 1980 to July 31st, 2015. (1) Literature were extracted according to inclusion and exclusion standards, and the quality of the extracted literature were evaluated by Newcastle-Ottawa Scale (NOS). (2) Meta-analysis was performed by RevMan 5 software formulated by using the Cochrane library databases. Various heterogeneity of the research was inspected firstly. According to the results of the inspection a certain effect model was selected (including fixed effects model, the random effects model) to be utilized in merger analysis. In this study pregnant women with both thyroid peroxidase antibodies (TPO-Ab) and thyroglobulin (TG-Ab) positive were defined as thyroid antibody positive pregnant women. Pregnant women with only TPO-Ab positive were defined as TPO-Ab positive pregnant women. Then the relationship of antibody positive and the risk of a pre-term birth was analyzed respectively. Results (1) Ten cohort studies were enrolled, of which NOS scale score were 7 or higher. All the studies are of medium quality and above. A total of 1 322 cases of preterm birth occurred among 19 910 pregnant women. (2) Positive thyroid autoantibodies did not increase the risk of preterm birth in euthyroid pregnant women (OR=1.41, 95%CI:0.83-2.40, P=0.200) or in pregnant women with hypothyroidism (OR=0.68, 95%CI: 0.32-1.44, P=0.310). Positive TPO-Ab in euthyroid pregnant women increase the risk of preterm birth significantly (OR=2.08, 95%CI:1.09-3.97, P=0.030), but positive TPO-Ab in pregnant women with hypothyroidism did not increase the risk of preterm birth significantly (OR=1.21, 95%CI:0.65-2.24, P=0.550). Conclusion Positive TPO-Ab is an independent risk factor of preterm birth in euthyroid pregnant women.