Objectives Mega cities carry a population of tens of millions, and accurately extracting the built-up areas of mega cities and conducting efficient settlement monitoring and risk analysis is crucial for ensuring the safety of urban residents and future sustainable development.

Methods First,combined with the Landsat series optical images and Sentinel-1 radar images, the maximum likelihood supervised classification method and time-series interferometric synthetic aperture radar （InSAR) technology were used to accurately extract the urban built-up area and settlement monitoring results of Guangzhou in 2017, 2020, and 2023. On this basis, an analytic hierarchy process model was established based on risk indicators including settlement velocity, uneven settlement rate, and density of building points in built-up area. Moreover, a long-term settlement risk assessment method for urban built-up area integrating spatiotemporal characteristics was proposed, by analyzing the cumulative changes of multi-temporal settlement and the spatiotemporal characteristics of point targets in depth. Finally, the settlement risk level of the built-up area in Guangzhou was quantitatively evaluated.

Results The extraction accuracy of built-up areas can reach over 91%, and their evolution is in line with the development trend of urban construction planning. Their spatial distribution is highly consistent with the distribution of InSAR points with a density greater than 3×10⁵ points/km². Among them, the settlement in Nansha and Baiyun areas is more significant, with the highest annual settlement velocity in some areas reaching -67 mm/a; the long-term settlement risk classification results of the research area indicate that about 1.69% of the areas have high risks and deserve attention.

Conclusions This study can provide data support for urban security decision-making and achieve rapid and low-cost regular risk screening of large-scale urban built-up areas.