Preterm birth is one of the main causes of perinatal morbidity and mortality, which does not tend to decrease in rate. The risk of death in premature babies is 2535 times higher than that of full-term babies, and stillbirths are registered 813 times more often than in timely delivery. To date, there are no effective ways to prevent preterm birth. Therefore, the timeliness of therapy, which largely determines the outcome of pregnancy in general, depends on the effectiveness of assessing the likelihood of their development. At the International Federation of Gynecology and Obstetrics (FIGO) Congress (2018), preterm birth is identified as a problem that has not yet been solved at the current stage of science and technology development. The result of the unsolved problems is a situation wherein the modern world over the past 60 years there has been no decrease in the premature birth rate, which is 9.5% of births and annually ends with the birth of 15,000,000 premature babies. The study aimed to research modern methods of diagnosis and prediction of spontaneous preterm birth. An analytical method was used in the study: a detailed systematic analysis of modern domestic and foreign literature on the diagnosis and prognosis of preterm birth. We used eLibrary, Scopus, PubMed, MEDLINE, ScienceDirect, Cochrane Library bibliographic databases (until August 2020). The article deals with the diagnosis and prediction of preterm birth probability, which will optimize the management of patients from the risk group and, in the future, will reduce the rate of perinatal morbidity and mortality of premature babies. Despite a significant number of researches devoted to the study of possibilities for diagnosing and predicting spontaneous preterm birth, currently, there are no methods with absolute diagnostic value. Most -existing studies indicate that when assessing the probability of preterm birth, a comprehensive approach should be preferred taking into account the results of several main and additional methods.