External corrosion is one of the main causes of equipment failures in industrial facilities, leading to highly costly maintenance. This study presents a machine learning approach to predict corrosion rates based on climatic data, area, and percentage of corrosion. The model employs the supervised learning algorithm Random Forest, leveraging a dataset of corrosion measurements collected over time. Additionally, the study aims to incorporate new data and variables into the model and evaluate their impact on prediction performance. Thus, the primary objective of this project is to enable prescriptive maintenance planning through the developed algorithm, ensuring operational safety and reducing costs.