Run-of-River Small Hydropower Plants (RoR SHPs) provide sustainable electricity but may threaten river ecosystems by altering flow regimes. This study develops a flexible framework to optimize RoR SHP operations, balancing energy generation with ecological preservation. Ten Environmental Flow Methods (EFMs) were assessed using Artificial Neural Networks (ANN), Response Surface Methodology (RSM), and a Multi-Criteria Decision-Making (MCDM) framework combining Entropy Weight Method and TOPSIS. ANN achieved the best accuracy (R2 = 0.9602; MAPE = 5.12 %), outperforming RSM (R2 = 0.9466; MAPE = 6.02 %) and a PCA-based regression benchmark (R2 = 0.9453; MAPE = 6.03 %). Among the scenarios, EFM7—allocating a minimum flow of 0.9  m3/s based on the lowest annual discharge—scored highest in TOPSIS (0.7650 for ANN; 0.7825 for RSM), achieving an effective balance between …