Spondylodiscitis is a serious infectious condition characterized by inflammation and destruction of the vertebral bodies and intervertebral discs, often leading to significant morbidity—including chronic pain, spinal deformity, and neurological compromise. In the Philippine healthcare setting, both tuberculous and non-tuberculous spondylodiscitis pose significant diagnostic and therapeutic challenges. An accurate diagnosis is essential to prevent unnecessary expenditures and decrease the burden on an already strained system. This pilot study determined if a Machine Learning (ML) model can distinguish between biopsy-proven tuberculous and non-tuberculous spondylodiscitis using Magnetic Resonance Imaging (MRI) features, with the goal of improving diagnostic accuracy for spondylodiscitis. This would enable timely and appropriate treatment to prevent complications such as spinal deformities or neurological deficits. A total of 28 patients seen at the Medical City Ortigas between January 2018 until May 2025 were screened and selected, including only those with laboratory or histopathologic confirmation of tuberculous or non-tuberculous spondylodiscitis. After analysis by board-certified radiologists, findings were systematically extracted based on established radiological criteria. A linear Support Vector Machine (SVM) algorithm was then used for model training. Results showed the efficacy of this model in differentiating tuberculous from non-tuberculous spondylodiscitis using MRI features, achieving an 85.7% accuracy and 0.93 AUC-ROC. On further analysis, imaging features were identified as key predictors such as abscess wall thickness, edge smoothness, and vertebral involvement. Findings in this study offer significant insight for both radiologists and orthopedic surgeons in facing the diagnostic challenge of differentiating between tuberculous and non-tuberculous spondylodiscitis. Although this study was limited by a small sample size, it can serve as a blueprint for further studies with a larger population to hopefully improve the accuracy and other key parameters of the machine learning model, which have both diagnostic and clinical implications.