In a world where energy is the lifeblood of diverse industries and daily life, achieving efficient and sustainable electricity production is paramount. This goal is exemplified by a prominent 350MW capacity power plant. One critical aspect of this endeavor is optimizing power plant startup processes, which include purging, firing, grid synchronization, temperature matching, and inlet pressure control. Delays in startup can result in penalties from the national grid. This report concentrates on the prediction of total startup time, a key element in enhancing power plant efficiency. To address this challenge, the project leverages big data tools and techniques. The goal is to analyze historical startup data patterns and develop a predictive model for estimating the time required for the power plant to be fully operational. The project unfolds in phases, beginning with data preprocessing and exploratory data analysis (EDA) to gain insights into the historical data’s patterns. Feature selection techniques are employed to identify the most relevant variables for predicting startup time. Machine learning algorithms, such as Random Forest Regressor, SVM Regressor, and Gradient Boosting Regressor, are then applied to create a predictive model. Notably, the Random Forest model consistently outperforms other algorithms, offering superior predictive accuracy. Evaluation metrics such as RMSE and R-squared are used to assess the model’s performance. The insights gained from this predictive model have the potential to revolutionize power plant operations by providing precise estimates of total startup time. By doing so, the project promotes dependable and efficient energy production by minimizing startup delays. This focused approach on predictive modeling for startup time estimation is crucial for improving the efficiency and sustainability of power plant operations in the energy sector.