Reliable fault detection, classification, and localization are critical for ensuring the stability and protection of high-voltage power transmission systems. This paper presents a unified multilabel framework that addresses these tasks simultaneously, providing a practical alternative to approaches that treat them independently. Using stratified evaluation to ensure balanced performance across fault types, the study demonstrates near-perfect accuracy and F1-scores in detecting and classifying faults, with equally strong performance in identifying faulted lines. Fault position localization, while more challenging due to overlapping feature distributions, achieves promising accuracy and highlights areas for further improvement. The proposed framework offers a scalable and efficient solution suitable for real-time deployment, contributing to enhanced grid reliability, reduced downtime, and improved operational efficiency in modern power systems.