Abstract: Pathogenic microbial aerosols (PMAs) are a major threat to human health. They are typically found in environmental pollutants and can cause a variety of health problems, including respiratory infections and pneumonia. In order to solve the main limitation that traditional HEPA system can only filter PMA, we developed a new plasma air purification system (PAPS), which can simultaneously filter PMA and kill internal microorganisms. The PAPS uses a large area needle corona discharge array to fully cover the airflow channel. This design allows for efficient filtration and disinfection of PMAs. The PAPS is also easy to clean and reuse, thanks to its proprietary modular design, and this makes it a low-cost option for long-term operation. In this study, artificial neural network and genetic algorithm (ANN-GA) are integrated to optimize the working parameters of PAPS. The reliability of the ANN-GA model is verified by effectively capturing and inactivating the PMA in the real world under the optimal process conditions. Moreover, the purification mechanism of PAPS system is studied by the multi-field coupled simulation of laminar flow field, electric field, plasma field and charged particle motion. The key working parameters are further verified. Experiments conducted under optimal working parameters show that the PAPS can effectively intercept and inactivate all bacteria in PMAs.