This study examines diabetes in pregnant women and the necessary parameters for timely diagnosis of gestational diabetes. Diabetes screening and diagnosis test is proposed for all pregnant women and for non-diabetic (normal and pre-diabetic) pregnant women. This research has used data mining methods to determine and apply the appropriate data mining algorithm and predict gestational diabetes in Iran. Available information can be accessed in the form of text files in the databases of public hospitals in the city of Ahvaz. The paper files of the patients of these hospitals are also used if needed. In these hospitals, the data of 8,882 pregnant women who visited midwives from 2013 to 2017 for prenatal tests are analyzed. This information is stored in the form of a database and completely coherent and separated in a similar system called Hospital information system (HIS) . This study has used various data mining algorithms such as decision tree for data analysis and also Rapidminer tool and Rstudio has been used for data analysis. According to the results, data mining software has correctly predicted 95.99% of the data, which indicates that it is a suitable method for predicting diabetes in pregnant women. According to the results obtained on the set of diabetic patients, among the classification methods, the decision tree obtains a better result, and K Nearest Neighbour also has a higher accuracy than the classification method, which can be placed as a support method along with doctors' decisions to improve the accuracy of predicting diabetes in pregnant women.
