The objective of the study was to investigate the role of prostaglandin D2 during pregnancy and its mediator Lipocalin-type prostaglandin D2 synthase (L-PGDS) as a predictor of preterm birth (PTB). Transgenic L-PGDS (+/+), L-PGDS (-/-) and C57BL/6 control pregnant mice models were used to determine the effect of DP1 and DP2 receptor antagonists in lipopolysaccharide (LPS)-induced PTB mice. In addition, L-PGDS levels were measured in the cervicovaginal secretions (CVS) of 370 pregnant women using ELISA and further processed for isoform detection using 2-D gel electrophoresis. Our results found that C57BL/6 control mice (n = 26), transgenic L-PGDS (+/+) (n = 26), demonstrated an 89% and 100% preterm birth in LPS (intraperitoneal injection, 20mg/kg) induced mice model respectively. Interestingly, the incidence of PTB was significantly reduced to 40% in L-PGDS (-/-) knockout mice (n = 26). DP1 and DP2 receptor antagonists (0.264 μg/day, dose of 0.1 μg/μl with the flow of 0.11 μl/h for 28 day using Alzet pumps) were used to investigate the effect in LPS-induced PTB in C57BL/6 mice and found 3.3-fold increase in viable pups after LPS-induction. In addition, L-PGDS levels were measured in CVS samples and found that PTB women (n = 296) had two-fold higher levels compared to full term births (n = 74) and established a significant inverse correlation between levels of L-PGDS and days to expected delivery by using 370 preterm birth CVS samples. Elevated L-PGDS levels in the CVS of women may be considered as a potential biomarker for PTB in future. Secondly, the use of DP1 and DP2 receptor antagonists may represent novel tocolytic agents for the treatment of PTB.