Paddy is one of the most important food crops in China. Due to the intensive planting in the surrounding of rivers and lakes, paddy is vulnerable to flooding stress. The research on predicting crop yield loss derived from flooding stress will help the adjustment of crop planting structure and the claims of agricultural insurance. The paper aimed to develop a method of estimating yield loss of paddy derived from flooding by multi-temporal HJ CCD images. At first, the water pixels after flooding were extracted, from which the water line (WL) of turbid water pixels was generated. Secondly, the water turbidity index (WTI) and perpendicular vegetation index (PVI) was defined and calculated. By analyzing the relation among WTI, PVI and paddy yield, the model of evaluating yield loss of flooding was developed. Based on this model, the spatial distribution of paddy yield loss derived from flooding was mapped in the study area. Results showed that the water turbidity index (WTI) could be used to monitor the sediment content of flood, which was closely related to the plant physiology and per unit area yield of paddy. The PVI was the good indicator of paddy yield with significant correlation (0.965). So the PVI could be used to estimate the per unit area yield before harvesting. The PVI and WTI had good linear relation, which could provide an effective, practical and feasible method for monitoring yield loss of waterlogged paddy.