Background: Neonatal mortality and neonatal morbidity are increased by preterm birth. Researchers have assessed various biochemical and biophysical markers to predict women at risk for preterm birth in order to decrease its incidence, prevent complications and improve survival rate for infants. Objective: To evaluate the prediction of spontaneous preterm birth by estimation of salivary progesterone levels and by serial measurements of cervical length. Patients and methods: A prospective study was carried out at the Maternity and Child Hospital, Basrah, Iraq. Symptomless women (n = 40) with a single fetus, and with a preterm birth, preterm rupture of membranes, or late spontaneous miscarriage (at 20–28 weeks of pregnancy), were recruited from the outpatient clinic at 24–28 weeks of gestation and were retested after 4 weeks. At each visit, three salivary samples were collected and assessed for their progesterone level using ELISA. Transvaginal sonography was used for cervical length estimation. Results: Of the 40 women, 9 were delivered at term (i.e., after 37 weeks). Their mean pregnancy duration at time of parturition was (38 \pm 2) weeks. The other 31 women underwent preterm birth, 13 cases with early preterm (\ge24–34 weeks) and 18 cases with late preterm (>34–37 weeks). The mean levels of salivary progesterone for the term group at the first visit (24–28 weeks) and the 2nd visit (28–34 weeks) were 797.2 pg/mL and 899.4 pg/mL respectively. The mean levels of salivary progesterone in the preterm group at the first visit and 2nd visit were 344.2 pg/mL and 257.3 pg/mL respectively. The difference between the term and preterm birth groups was statistically significant (P = 0.04). There were significant differences regarding cervical length measurement between the preterm delivery groups and the term delivery group. Conclusion: Both low salivary progesterone concentrations and low cervical length can be applied for predicting preterm birth in asymptomatic at-risk women.