Industrial pollutants are a major contributor to urban air pollution, negatively impacting environmental sustainability and public health. To address this challenge, we propose a novel LSTM-Random Forest-Based Industrial Emission Control System that integrates the sequential modeling capabilities of LSTM networks with the robust interaction analysis of Random Forest algorithms. The system collects real-time data on key pollutants such as \(PM_{2.5}\) , \(NO_{2}\) , and CO, along with climatic variables including weather conditions, temperature, wind speed, humidity, and pressure, via IoT sensors positioned near industrial sources. The LSTM model forecasts emission patterns over both short and long-term horizons by capturing temporal dependencies in the data, while the Random Forest algorithm enhances prediction accuracy by incorporating contextual factors like industrial activity levels and weather conditions. Experimental evaluations demonstrate that our hybrid approach achieves more accurate and reliable forecasts than traditional methods, enabling dynamic adjustments in industrial operations to optimize production, activate emission control systems, and ensure regulatory compliance. This integrated strategy not only offers a flexible framework for reducing industrial emissions and promoting sustainable practices but also lays the groundwork for future enhancements such as comparisons with state-of-the-art models and the adoption of edge computing for real-time decision-making.