One of the most common and severe complications of premature birth is bronchopulmonary dysplasia (BPD). Early intervention and avoidance of subsequent detrimental effects depend on accurate and timely diagnosis using prediction tools. The purpose of this research was to use machine learning in the theory of BPD’s developmental etiology to create a tool for predicting whether or not a person may acquire BPD. The first models were developed using datasets consisting of perinatal variables and early postnatal respiratory support; then, the two models were combined into a final ensemble model by means of logistic regression. Clinical scenarios were simulated. The study participants totaled 689 newborns. To create the model, researchers collected information from 80% of newborns, while the remaining 20% was used for validation. Receiving operating characteristic curves were used to evaluate the final model’s performance, and they revealed values of 0.921 (95% CI: 0.899-0.943) for the training dataset and 0.899 (95% CI: 0.848-0.949) for the validation dataset. Extubation to CPAP has been shown to be more effective than NIPPV in preventing brain perfusion deficits in computer models. The absence of the need for reintubation within 9 days of the first extubation can also be considered evidence of a favorable outcome following tracheostomy. Prediction of BPD using machine learning to analyze perinatal characteristics and respiratory data may be clinically useful for facilitating early targeted care in high-risk infants.