Water quality degradation poses a critical challenge to the sustainable use of global freshwater resources, especially in semi-arid regions where water scarcity is severe. However, limited attention has been paid to water quality and critical pollutants in semi-arid rivers, and effective methods of water quality evaluation are deficient in these vulnerable regions. This study conducted an integrated water quality evaluation by combining the single-factor (SF) and water quality index (WQI) methods in two large semi-arid Eurasian rivers (i.e., the Irtysh River and Ili River) between China, Kazakhstan and Russia. Meanwhile, the critical pollutants were identified based on the WQI and a multivariate statistical method. The integrated evaluation indicated overall "good" water quality, yet revealed potential hazards from elevated total suspended solids (TSS), nitrogen and organic pollution. Water quality degraded markedly during the flood season, particularly in the middle reach of the Irtysh River and the lower reach of the Ili River. TSS was identified as the dominant driver of water quality degradation, suggesting that sediment-related physical impairment can outweigh nutrient signals in human-impacted semiarid rivers. An optimal minimal water quality index (WQImin) model, consisting of TSS, permanganate index (CODMn), nitrate (NO3-N), ammonium (NH4-N), and water temperature (WT), achieved improved predictive accuracy when expertise-based weights were applied. The WQImin model also remained effective when transferred between the two hydrogeomorphologically comparable basins. These findings support WQImin as a costeffective option for routine monitoring in data-limited, drought-prone regions and suggest that management should prioritize sediment control and the inclusion of TSS in regulatory monitoring frameworks, alongside watershed-scale erosion and runoff mitigation.