The rapid detection of methanol is essential for public health, particularly in preventing methanol poisoning, which can lead to irreversible visual damage or death. Conventional laboratory-based analytical methods are often too slow for timely intervention during exposure incidents, highlighting the need for rapid, low-cost screening technologies. In this work, we present a chemiresistive gas sensor array based on Ti₃ C₂ T_x MXene, synthesized via an in-situ HF etching pathway and deposited onto interdigitated electrodes via dielectrophoresis to establish a highly conductive sensing layer. To address the inherent selectivity challenges associated with chemically similar volatile organic compounds, a machine-learningassisted data analysis pipeline is implemented, extracting 19 features from the transient sensor responses. Principal component analysis and linear discriminant analysis provided clear class separation of methanol from ethanol and acetone. The trained classifier reached an accuracy of 99% for same-day measurements and retained an accuracy of 83% for data acquired on subsequent days, demonstrating robustness and generalizability of proposed algorithms against sensor temporal drift. These results highlight the potential of machine learningenabled Ti₃ C₂ T_x MXene-based 'electronic nose' systems as portable tools for rapid methanol screening and practical VOC discrimination.