River discharge is an important parameter in understanding water cycles, and consistent long-term discharge records are necessary for related research. In practice, discharge records based on in situ measurement are often limited because of technological, economic, and institutional obstacles. Satellite remote sensing provides an attractive alternative way to measure river discharge by constructing an empirical rating curve between the parameter provided by remote sensing techniques and simultaneous ground discharge data. River width is a popular parameter for constructing the empirical curve, since change in river discharge can be represented by a change in river width. In some rectangular channels, however, river width does not change significantly with river discharge, so an alternative parameter is necessary. We analyze a novel technique using river island area as an indicator of discharge. A river island often has a flat terrain, and its area decreases with higher discharge. This technique is validated by three river islands in the Yangtze River in China. All 61 remotely sensed images acquired by the HuanJing (HJ) satellites from 2009 to 2010 were correlated with corresponding in situ discharge of the nearby Zhicheng hydrological station. The performance of fitted curves for inferring river discharge is validated using 36 HJ images taken in 2011, and the influence of remotely sensed imagery and river islands is discussed. All three river islands can be used as indicators of river discharge, although their performances are much different. For the river island with the best result, the mean accuracy of the estimates is less than 10% of the observed discharge, and all relative errors are within 20%, validating the effectiveness of the proposed method.