QUiPP, developed by researchers at King's College London, was tested in two studies of high-risk women being monitored at ante-natal clinics.
The app uses an algorithm which combines the gestation of previous pregnancies and the length of the cervix with levels of foetal fibronectin to classify a woman's risk.
The first study focused on women deemed to be at high risk of preterm birth, usually because of a previous early pregnancy, despite not showing any symptoms.
The second study predicted the likelihood of early delivery in a group of women showing symptoms of early labour which often does not progress to real labour.
More From This Section
The estimated probability of delivery before 30, 34 or 37 weeks' gestation and within two or four weeks of testing for foetal fibronectin was calculated for each patient and analysed
as a predictive test for the actual occurrence of each event.
In the second study, data from 382 high-risk women was collected. The model was developed on the first 190 women and validated on the remaining 192.
Worldwide 15 million babies are born preterm (before 37 weeks) each year and over a million of these die of prematurity-related complications, researchers said.
A number of factors are used to determine if a woman is at risk of giving birth prematurely, including a history of preterm births or late miscarriages.
Two further factors which doctors can consider are the length of cervix and levels of a biomarker found in vaginal fluid known as foetal fibronectin, which are typically tested from 23 weeks.
"Despite advances in prenatal care the rate of preterm birth has never been higher in recent years, including in the US and UK, so doctors need reliable ways of predicting whether a woman is at risk of giving birth early," said Andrew Shennan from King's College.
"It can be difficult to accurately assess a woman's risk, given that many women who show symptoms of preterm labour do not go on to deliver early," he added.
The findings were published in the journal Ultrasound in Obstetrics & Gynecology.