Electric vehicles (EVs) are a promising solution for sustainable transportation. EVs rely on lithium-ion batteries, and their performance is monitored by the Battery Management System (BMS). The State of Health (SOH) is a critical parameter that indicates the battery's health over time. Accurate diagnosis of battery states, including SOH, is crucial for safety and preventing failures. The paper proposes an Optimized Long Short Term Memory (LSTM) model based on Bayesian Hyper-parameter Optimization. LSTM is chosen over other methods due to its unique advantages. Unlike conventional estimation approaches like regression or statistical models, LSTM excels in handling long-term dependencies and capturing intricate temporal relationships within battery data. Its architecture enables the retention and utilization of historical information, making it highly suitable for analyzing dynamic and evolving battery behaviours. Unlike existing approaches, the Optimized LSTM model demonstrates superior performance and high accuracy, making it practical for real-world applications.