In Industry 4.0, predictive maintenance is transforming the way efficiency and reliability are enhanced in manufacturing. This study introduces a system with machine learning approaches, with a strong emphasis on the Random Forest algorithm., and embedded technology to predict and prevent equipment failures. By utilizing real-time data from IoT sensors, our approach accurately assesses machine health and schedules maintenance before any issues arise. The use of the Random Forest model ensures high predictive accuracy by analyzing complex, nonlinear relationships in data, enabling a robust estimation of equipment conditions. This proactive strategy significantly reduces unexpected downtime, lowers maintenance costs, and prolongs machinery lifespan. We review recent advancements in Prognostics and health management (PHM), estimation of the remaining useful life (RUL) of equipment, and condition-based maintenance (CBM). Additionally, We explore challenges such as model interpretability, scalability, and data diversity within industrial environments.