Abstract: The Qinghai-Tibet Plateau region is rich in water resources. Flood magnitude analysis for different recurrence periods in the context of climate change can provide scientific support for the construction of local water conservancy projects; however, related research remains limited. This study uses the Parlung Zangbo River basin as a case study, integrating the distributed hydrological model VIC-Glacier with interpretable machine-learning methods. Data from the sixth phase of the Coupled Model Intercomparison Project （CMIP6） and projected glacier change results revealed the future flood magnitudes for near-term （2027—2040）, mid-term （2047—2060）, and long-term （2087—2100） periods under both moderate （SSP2-4.5） and high （SSP5-8.5） forcing scenarios for different recurrence periods. The results discovered major spatial heterogeneity in flood magnitudes and their change characteristics within the Parlung Zangbo River basin. Under future climate scenarios, decline in glacier coverage and increased precipitation will alter runoff components. In glacier-covered basins, future flood magnitudes for different recurrence periods decreased by 6.46%—32.27%. Conversely, in non-glacier basins, flood magnitudes for different recurrence periods showed a rising trend, with increases of 2.87%—15.52%.