Industrial activities have released liquid droplets, solid particles, and gas molecules into the atmosphere, with high concentrations of particulate matter (PM2.5) posing serious health risks, making the assessment of particulate matter concentration crucial for enhancing human well-being and aligning with Industry 5.0's focus on sustainable and human-centric development. To tackle this issue, this study proposes a novel ensemble model based on artificial neural networks and boosting regressors (XGBoost, CatBoost, and LightGBM) to forecast PM2.5 levels using data from eight cities in Beijing, China, spanning from 2010 to 2017. The ensemble model utilizes the bootstrap aggregating method and is evaluated with Root Mean Square Error (RMSE), Mean Square Error (MSE), and the Coefficient of Determination (R²). The model's performance surpasses existing models, achieving a significantly lower RMSE of 0.076 and an impressive R² of 0.999, indicating high accuracy. Estimating PM2.5 concentrations is crucial for sustainable industry, enabling effective air quality management, regulatory compliance, public health protection, and cleaner production practices.