Objectives: Prenatal health is an important factor in the wellbeing of the mother and fetus. One prenatal wellbeing factor has been determined as psychological flexibility that enables individuals to adapt to the evolving situation and cope with stress. Objectives: The purpose of this research is to test the mediation of the maternal-fetal attachment between psychological flexibility and prenatal wellbeing relationship. Methods: Data collection was carried out to 300 pregnant women at three pregnancy stages (first, second, and third trimesters) by adopting a mixed-methods design. The data was pre-processed using Min-Max Normalization and then Recursive Feature Elimination was used to determine the most relevant predictors. To predict prenatal wellbeing, a Random Forest Classifier was applied and the performance of the model was measured on accuracy and feature importance. It was analysed using Python, scikit-learn to perform machine learning, manipulate data using pandas, and statistically model using stats models. Results: The Random Forest Classifier has an accuracy of 87% in predicting the prenatal wellbeing, which means that it fits well to the model. The main predictive variables were psychological flexibility (particularly, mindfulness and emotional regulation) and maternal fetal attachment, which was reported to mediate the correlation between psychological flexibility and prenatal wellbeing. The statistical tests showed that psychological flexibility was statistically significant in the direct enhancement of prenatal wellbeing (p<0.01) and the indirect enhancement via maternal fetal attachment (p<0.05). Conclusion: The results emphasize the importance of psychological flexibility in enhancing the prenatal wellbeing, and maternal-fetal attachment is one of the main mediators. The high level of accuracy of the model implies that psychological flexibility-based interventions may be effective in improving maternal health and wellbeing in pregnancy.