ABSTRACT On-Load Tap Changers (OLTCs) are essential components in oil transformers, ensuring stable voltage regulation under varying load conditions. Traditional maintenance methods are often reactive, causing unplanned failures, operational downtime, and substantial repair costs. This study proposes a predictive maintenance framework that utilizes Dissolved Gas Analysis (DGA) to detect early signs of OLTC failure. Data from transformers operating at 33/11 kV were analyzed using DGA, the Duval Triangle method, and Failure Mode and Effect Analysis (FMEA). The results show a marked reduction in fault gas concentrations, improved dielectric breakdown voltage (BDV), and a significant shift from high risk fault zones to safe operational conditions. This predictive approach demonstrates enhanced maintenance efficiency, reduced downtime, and an extended transformer lifespan, providing a proactive solution for improving transformer reliability.