Road safety and transportation system operation mainly depends on sign boards. The proposed system presented in this paper uses the Internet of Things (IoT) and random forest regression to improve road sign maintenance. Insufficient resource allocation results from timetabies and manual inspections in traditional maintenance techniques. The suggested approach uses IoT sensors mounted on road signs to collect real-time weather, visibility, and traffic volume data. This data helps us estimate road sign degradation over time, improving maintenance accuracy and cost-effectiveness. Random forest regression is essential since it handles complicated, non-linear data connections well. It trains the model using historical information encompassing varied elements affecting road sign deterioration, such as sunlight and moisture exposure, to construct predictive models that can reliably estimate maintenance requirements. The comprehensive experiments and real-world testing show that random forest regression reliably predicts road sign deterioration better than existing techniques. The algorithm's ensemble nature improves forecast accuracy, while feature significance analysis identifies degradation dri verse © 2024 Elsevier B.V., All rights reserved.