This research paper compares the performance of Naive Bayes, Artificial Neural Networks (ANN), and Decision Tree algorithms in detecting Borderline Personality Disorder (BPD) using two datasets. Accuracy, F1-score, recall, and precision metrics are calculated to evaluate BPD detection effectiveness. The datasets, derived from clinical records and an online mental health platform, undergo preprocessing techniques for data integrity and comparability. Naive Bayes, ANN, and Decision Tree algorithms are trained and tested independently on both datasets, and their performance is assessed using accuracy, F1-score, recall, and precision metrics. Results reveal varying performance across the datasets. Naive Bayes achieves higher accuracy on Dataset A, while ANN and Decision Tree perform better on Dataset B. This research provides insights into dataset characteristics' impact on BPD detection performance, emphasizing the importance of algorithm and dataset selection. Findings guide clinicians, researchers, and developers in selecting suitable algorithms and datasets for accurate BPD detection and intervention strategies.