<p>Land use/land cover changes (LULCC) information serves as an essential foundation for understanding urban development and implementing territorial spatial planning. Due to their unique geographical environments, complex spatial structures and high landscape fragmentation, cities with dense waterway networks have received limited attention in long-term and systematic research on LULCC. Based on Landsat time-series imagery, this study developed the LULCC information extraction model for cities with dense waterway networks, utilizing Google Earth Engine (GEE) and a Random Forest classification method that integrates multi-feature parameters. Experimental results in Suzhou show that: (1) The Random Forest classification algorithm, which integrates multi-feature parameters, accurately extracts LULCC information in Suzhou, achieving a mean Overall Accuracy (OA) of 86.17% and a Kappa coefficient of 0.83. (2) During the study period, the area of construction land (from 305.23 km² to 2633.29 km²) and other land (from 0.15 km² to 0.18 km²) showed an increasing trend. While the areas of cropland/grassland (from 5426.99 km² to 3537.26 km²), water bodies (from 2725.82 km² to 2312.49 km²), and forest land (from 171.22 km² to 146.19 km²) showed a decreasing trend. Land use transition was dominated by the conversion of cropland/grassland to construction land, with the transition intensity strengthening over time. (3) The center of gravity for water bodies generally shifted southwest, showing a convergence towards ecological core areas such as Taihu Lake; the center of gravity for construction land continuously migrated southeast, confirming the strategic layout of eastward urban development. The centroid of cropland/grassland exhibited oscillatory fluctuations, while forest land and other land showed significant transient disturbances, revealing the dynamic adjustment and response mechanisms of land cover patterns under high-intensity human interference. (4) Landscape indices analysis indicates that landscape heterogeneity and diversity in Suzhou have significantly increased, with an overall evolutionary characteristic of intensified fragmentation and decreased connectivity. However, driven by intensive territorial spatial management and control since 2015, the fragmentation trend has been initially mitigated, and the landscape structure has shown a trend of local integration. The integrated framework significantly enhances the identification of fragmented landscapes and mitigates spectral confusion, and provides a scientific basis for urban territorial spatial planning and regional sustainable development.</p> Graphical Abstract <p></p>