Abstract: A scientific understanding of the evolution patterns and influencing factors of hydro-ecological space in the urban agglomeration in the middle reaches of the Yangtze River is essential for the coordinated planning of regional water resource development and protection. This is also important for supporting the high-quality and sustainable development of territorial space. Based on annual time-series remote sensing data, this study integrated spatio-temporal cube analysis, Mantel test, and interpretable machine learning models to comprehensively examine the evolution patterns and the driving mechanisms of hydro-ecological space in the region. The results revealed that from 2000 to 2023, the hydro-ecological space decreased by 3,300 km2, exhibiting a spatio-temporal trend of “more in plains, less in mountains” and a shift of “decrease in plains, increase in mountains.” The evolution pattern displayed an accumulative effect with increasing temporal span, i.e. the cold and hot spot clustering intensifying over the long term. Structural transformation primarily occurred between hydro-ecological space and both agricultural and urban space, with the expansion of these two land types leading to a reduction of 2,765.30 km2 and 616.14 km2 in hydro-ecological space. The spatio-temporal evolution of hydro-ecological space in the urban agglomeration was driven by multiple interacting and nonlinear factors across the dimensions of natural geography, transportation location, and socio-economic conditions. To address the challenges posed by future climate change and the functional demands for hydro-ecological services in territorial spatial development and conservation, it is imperative to enhance governance capacity by aligning with the territorial spatial planning system, e.g.particularly through delineating hydro-ecological redlines, optimizing spatial layout, and establishing a collaborative governance framework.