Predicting diamond prices is a challenging task due to the complex and subjective nature of the diamond market. While ensemble methods have been widely used for predictive modeling, they often fail to provide accurate and reliable results. This study examines the potential of Bayesian additive regression trees (BART), a probabilistic model that combines likelihood and prior distributions for predictions. This research aims to implement an accurate predictive model for diamond price forecasting using BART and compare its performance with popular ensemble methods, such as random forest, gradient boosting machine, extreme gradient boosting, and light gradient boosting. This study utilized differently structured datasets, including the Diamond Financial Index and diamond-cut data, to evaluate the models’ performance. The analysis revealed that the BART regressor exhibited exceptional performance, achieving the highest R-squared \((R^2)\) score of 0.981 on the Diamond Financial Index and 0.984 on the diamond-cut data, outperforming the other ensemble models. The study’s findings suggest that the BART model is an effective tool for the diamond industry, providing a more accurate and reliable way to predict diamond prices. Furthermore, the research emphasizes the importance of considering different approaches, such as Bayesian modeling, in developing accurate predictive models, particularly in complex and subjective domains like the diamond market. Overall, this study provides valuable insights into the application of Bayesian modeling techniques for diamond price prediction and highlights the potential of the BART model as a powerful alternative to traditional ensemble methods.