<p>This study presents a novel ensemble learning approach for optimizing a fuzzy inference system (FIS) to forecast PM₁₀ pollution levels. The proposed model integrates the bagging technique with a fuzzy inference system to enhance predictive accuracy and robustness. By creating an ensemble of fuzzy inference systems, each trained on different subsets of the data, the approach leverages model diversity to improve overall performance. Optimization algorithms are employed to fine-tune the FIS parameters, further enhancing the model’s predictive capabilities. The effectiveness of the optimized bagging fuzzy inference system is evaluated using a real-world dataset of PM₁₀ pollution levels. Results indicate that the proposed method outperforms the traditional bagging algorithm, including the widely-used Random Forest, regarding accuracy and robustness. The optimized bagging fuzzy inference system offers a promising solution for reliable air quality forecasting, demonstrating its potential as a valuable tool for environmental monitoring and public health protection.</p>