In this research, real-time data and machine learning are used to construct a prediction system for agricultural pest infestation. In order to provide early pest infestation forecasts, the system combines a random forest classifier with a soil moisture sensing mechanism to analyse temperature, weather, and soil moisture levels. The model is trained using a dataset of crop-specific infestations, appropriate weather, and pest kinds. Users may enter crop characteristics on the platform, which was developed using the MERN stack, and get forecasts about potential pest infestations along with the associated probability. The technology uses real-time data from sensors and meteorological APIs to assist farmers in preventing insect outbreaks and enhancing crop output and health. With an emphasis on usability and prediction accuracy, the project provides a scalable, data-driven solution to a significant agricultural_problem. In order to ensure that the projections are updated with the most recent information, the system also uses a Raspberry Pi to collect real-time environmental data, such as soil moisture and weather conditions. The random forest classifier has been optimised to optimise predicted accuracy because of its resilience in managing big datasets and intricate decision-making. With the help of the MERN stack, the web interface offers farmers an intuitive platform for entering crop information and receiving real-time notifications of pest infestations. This initiative intends to reduce crop damage, maximise pesticide usage, and eventually support more sustainable farming methods by automating the prediction process and providing data-driven insights.