Anthropogenic impacts and climate change modify instream flow, altering ecosystem services and impacting on aquatic ecosystems. Alpine rivers and streams on the Qinghai-Tibet Plateau (QTP), are especially vulnerable to disturbance due to a limited taxonomic complexity. The effects of variations in flow have been studied using specific taxa, however, the flow-biota relationships of assemblages are poorly understood. A multi-metric habitat suitability model (MM-HSM) was developed, using biological integrity measures of macroinvertebrate assemblages to substitute for habitat suitability indices (HSI) derived from individual taxa. The MM-HSM was trained using macroinvertebrate data from three representative alpine rivers (the Yarlung Tsangpo, the Nujiang, and the Bai Rivers) on the QTP, and was verified using data from the Lanmucuo River. The model produced reliable predictions using the training dataset (R2 = 0.587) and the verification dataset (R2 = 0.489), and was robust to inter-basin differences and changes in dataset size. By coupling the MM-HSM with hydrodynamic simulations, the relationship between weighted usable area (WUA) and flow variations (0.11–1.99 m3/s) for macroinvertebrates was established, and a unimodal response pattern (optimal flow Q = 1.21 m3/s) was observed for macroinvertebrate assemblages from the Lanmucuo River. This was in contrast to the skewed unimodal or monotonically increasing relationships observed for individual indicator taxa, supporting our hypothesis that biological integrity varies with changing flow and conforms to the intermediate disturbance hypothesis. The MM-HSM provides a novel framework to quantify species-environment relationships, which may be used for integrated river basin management.