Sedighkia, MahdiDatta, Bithin2025-05-232025-05-232073-4441ORCID:/0000-0003-1730-8928/work/184099480http://www.scopus.com/inward/record.url?scp=85218990687&partnerID=8YFLogxKhttps://hdl.handle.net/1885/733752084This study presents a simulation–optimization framework that integrates deficit irrigation strategies with ecological considerations to mitigate the impact of water abstraction on potential fish populations in river ecosystems. The framework addresses two primary objectives: minimizing fish population loss, an ecological index reflecting environmental impacts, and minimizing the yield reduction of rice crops caused by deficit irrigation. Regression models and adaptive neuro-fuzzy inference systems were employed to simulate the physical and water quality parameters of the river. Additionally, a multivariate linear regression model was developed to estimate potential fish populations using combined physical and water quality indices as inputs. Multi-objective particle swarm optimization was applied to achieve the defined objectives. Results from the case study demonstrate the model’s ability to balance ecological requirements with rice production through deficit irrigation. The ecological degradation of river ecosystems was found to be comparable during dry and normal years, while rice yield decreased by approximately 10% in dry years. Comparisons with unsustainable practices, where ecological flow was disregarded, revealed that significant reductions in rice production are inevitable to sustain river ecosystems. The proposed method provides a practical approach for achieving a fair balance between agricultural benefits and environmental sustainability in river basins, making it a valuable tool for water resource management.17en© 2025 The Author(s)data-driven modelecological degradationsmulti-objective optimizationrice yieldriver ecosystems202510.3390/w1704053585218990687