Abstract Introduction Infants diagnosed with Congenital Hypothyroidism (CH) can have normal or near-normal cognitive abilities if their disease is diagnosed at an early stage. The newborn screening program plays a crucial role in the diagnosis and treatment of infants having CH. However, human error and lack of a proper method for detailed interpretation of the data gathered by the screening programs affect the success rate and consequently, the cognitive outcomes. Data mining techniques can notably assist experts in diagnosing CH or in choosing the best medication. Methods This cross-sectional study investigates the Support Vector Machine (SVM), Multilayer Perceptron (MLP), Chi-Squared Automatic Interaction Detector (CHAID), and Iterative Dichotomiser-3 (ID3) for diagnosis of CH. The newborn screening data of 4812 infants, registered in the Health Center of Alborz province, Iran, in 2016, was used to train and test the corresponding classifiers. Since this dataset is imbalanced, we combined the mentioned classifiers with Bootstrap aggregating (Bagging) and boosting techniques, to avoid the negative effects of this problem on the classification results and to achieve more accurate classifiers. Results The SVM-Bagging technique had the best performance, with precision and specificity of 100%, recall of 73.33%, F-measure of 84.62%, and an accuracy of 99.58%. Results show acceptable capabilities of this classifier to analyze newborn screening data. The overall results also demonstrate that besides blood test results, other factors such as gender, type of marriage, type of delivery, family history of thyroid disease, and chronic diseases, contributed in identifying infants with a high risk of CH. Conclusions Results showed acceptable capabilities to analyze CH screening data for the SVM-Bagging technique. This method can be considered as a diagnostic tool to assist experts in recognizing infants suspected of having the disease, and can increase the effectiveness and accuracy of diagnosis and treatment of CH at an early stage. This also reduces costs of postnatal therapy as compared with CH when diagnosed at later stages.