The study aimed to determine the most useful model for predicting functional constipation (FC) among college-bound girls in Kolkata by evaluating the applicability of multiple models and assessing the forecasting accuracy of prediction methods, such as regression-based and machine learning models.The observational descriptive study involved 300 college girls aged 18–25 from Kolkata, randomly selected via social media. Data were collected through an online questionnaire, and 19 attributes were selected for the prediction study. Weka version 3.8.0 software was used for predictive modeling, performance analysis, and building an FC prediction system. The data were divided into 70% training and 30% test datasets for each investigation. The results showed that 96.00% of instances were correctly classified, with a Kappa value of 0.875, a root mean squared error of 0.19, and an accuracy of 96.3%. The model achieved a weighted precision of 96%, 96% true positives, 0.05% false positives, an F-measure of 0.961, and an ROC curve of 0.994. Surprisingly, all six evaluators predicted Bristol's Stool Consistency Scale as the number one predictor of FC among college-going girls, followed by "Pain and discomfort in the abdomen" as the second predictor. In conclusion, this machine learning model-based automated approach for predicting functional constipation can assist medical professionals in identifying younger individuals more likely to experience constipation.