The scale of satellite images affects the recognition of landscape patterns and conditions. Therefore, this research focused on relationships between landscape fragmentation and function using multi-sensor imagery in semi-arid regions of Iran. Landscape function leakiness index (LI) and fragmentation metrics were calculated based on Sentinel-2, Landsat OLI, and MODIS data and compared using regression models and principal component analysis. Upscaled data were obtained based on changing the resolution of native Sentinel-2 from 10 to 30 m and 250 m. Results showed low (4%) and high (91%) fragmentation values in good and poor rangelands, respectively. The Sentinel-2 native LI index was the best indicator of vegetation cover (R² = 0.99, p < 0.001) and fragmentation (R² = 0.72, p < 0.05) and its upscaling to 30 m correlated better with fragmentation values than 250 m. The findings indicated that landscape functionality has reduced in the degraded areas because of increasing fragmentation and inter-patch spaces.