The machine learning (ML) as an indispensable tool has attracted widespread attention in domestic wastewater treatment processes.However, ML have not been widely utilized in comprehensive prediction of effluent quality and comprehensive environmental impact.In this research, under six ML models, 360 samples were collected to predict effluent quality and comprehensive environmental assessment (CEA) for anaerobic-anoxic-oxic-membrane bioreactor (A2O-MBR) domestic wastewater treatment process.The typical industrial scenario anal. indicated that the environmental impact could be decreased under the hydraulic retention time of 4-10 h, and nitrox reflux ratio in 50 %-200 %, when the influent total phosphorus in 1-2 mg/L, influent COD in 100-120 mg/L, influent ammonium in 10-18 mg/L.Meanwhile, a comprehensive index prediction model was developed, demonstrating that the extreme gradient boosting model has better fitting performance coefficient of determination with 0.8053.This research offers the foundation for the efficient and stable operation and environmental impacts of the wastewater treatment process.