Urban blue-green space cooling island effect (BGCI) is effective in improving the thermal comfort of residents. However, there is little knowledge regarding the diurnal variation of BGCIs and the influencing factors. Therefore, we selected Beijing as the study area and used ECOSTRESS LST data and the inflection–maximum perspective method to explore the diurnal variation of BGCIs. Additionally, we investigated diurnal variations in the relative influence of the characteristics of the blue-green space itself, as well as the surrounding 2D and 3D landscape metrics using boosted regression tree model. The results indicated that BGCIs displayed distinct diurnal patterns. BGCIs progressively increased from sunrise to midday, decreased thereafter to sunset, reached its peak around midday, and diminished to a relatively low level and constant intensity at night. BGCIs of water bodies exhibited a significantly higher intensity compared to vegetation during the day, particularly around midday, with a difference in mean cooling intensity (CI) of 1.06 °C and mean cooling distance (CD) of 63.27 m, while the differences were minimal at night with a difference in mean CI of 0.02 °C and mean CD of 9.64 m. The features of vegetation had a more significant impact on BGCIs during the day, particularly around midday (CI: 32.30% around midday and 13.86% at night), while the 3D metrics influenced BGCIs more at night (CI: 26.40% around midday and 35.81% at night). The features of water bodies had a greater impact during the midday (52.87% around midday and 10.46% at night), with the landscape metrics of surrounding water bodies playing a more important role at night (15.56% around midday and 38.28% at night). The effect of tree height, shape index of vegetation, and surrounding building coverage ratio of water bodies on BGCIs exhibited opposite trends around midday and at night. Optimizing the landscape surrounding blue-green spaces is more cost-effective than the blue-green spaces themselves for nighttime thermal comfort, especially in 3D urban landscapes. These findings emphasize the imperative and essentiality of exploring diurnal variations in BGCIs, providing valuable information for mitigating UHI effects.
