Hydrological connectivity (HC) is a useful framework for understanding hydrological responses to landscape changes. We present herein a novel model (SOFAR) for utility-scale solar farms (USFs), combining modules of soil moisture dynamics, roof effects of photovoltaic panels (PVs), vegetation growth and landform evolution. By augmenting the model with a DEM-based HC index, we investigate hydrological behaviors following the construction of a USF in China's Loess Hilly Region. Nine scenarios are designed, to explore the effects of co-evolving ecohydrology and landscape on soil erosion and HC in USFs deployed in different climates and terrains, by altering the annual precipitation, rainfall frequency, and ground slope. Our results show that the USF considerably increased runoff (99.18%–154.26%) during its operational period, and soil erosion rate (21.4%–74.84% and 25.35%–76.18%) and HC (0.08%–0.26% and 0.47%–0.91%) throughout construction and operational periods, respectively. The highest erosion rates were detected in the PV installation zones and in the areas close to the river channel. We prove the hypothesis that HC is a critical indicator for sediment yield in a USF, and thus the long-term responses of soil erosion to USF installation and development can be explained in terms of HC. We conclude that USFs may increase soil erosion, mainly by increasing local HC and runoff, and higher background HC may in turn further aggravate the effects of USFs on soil erosion. Our results underscore the importance of including landscape ecohydrologic and geomorphic feedbacks, to improve the environmental impact assessment of USFs.