<p>Image velocimetry has gained significant attention for flow measurement in field rivers, with recent efforts combining satellite videos with various image velocimetry techniques to enhance cost-effectiveness, resolution, and public accessibility. However, challenges remain, including lower image resolution and frame rates compared to conventional methods, as well as pixel offsets caused by continuous satellite motion. Moreover, the accuracy of satellite-based flow measurement has yet to be systematically evaluated due to the difficulty of obtaining instantaneous and precise surface flow fields in large rivers. This study applied three mainstream image velocimetry methods, namely space-time volume velocimetry (STVV), optical flow methods (OFM), and large-scale particle image velocimetry (LSPIV), to analyze satellite video footage of the Chongqing section of the Yangtze River. A hydrodynamic simulation was conducted to generate reference instantaneous flow fields corresponding to the video acquisition date. Results indicated that all three methods can derive large-scale flow fields with acceptable accuracy (the root mean square errors of velocity magnitude for STVV, OFM, and LSPIV were 0.3218 m/s, 0.8131 m/s, and 0.8003 m/s) when measurement parameters were appropriately adjusted, demonstrating the feasibility of satellite-based river velocity monitoring. Among them, STVV achieved the best performance, with regression lines closely aligned with y=x. The coefficients of determination were 0.86 for flow direction and 0.82 for velocity magnitude, with a relative error of 14.24% of the mean simulated velocity. Further analysis revealed that image resolution, frame rate, and STVV-specific calculation parameters influenced measurement accuracy, and a derived relationship between measurement error and key parameters can guide parameter selection.</p>