Abstract: In order to investigate the effects of waterlogging stress during jointing and tasseling stages on the grain filling of summer maize, field experiments were conducted during summer maize growth seasons in 2011 and 2012. ‘Xundan20’, a widely cultivated maize variety in the Huang-Huai Plain, was used in the study. Ten treatments were set up at jointing and tasseling stages, including surface waterlogging for 3 or 5 days (JF3 or JF5), and subsurface waterlogging for 5, 7 or 10 days (JW5, JW7 or JW10) at jointing stage; and surface waterlogging for 3 or 5 days (TF3 or TF5), and subsurface waterlogging for 5, 7 or 10 days (TW5, TW7 or TW10) at tasseling stage. In the surface waterlogging treatments, 5 cm depth of water layer was maintained on the soil surface. In the subsurface waterlogging treatments, soil water content was maintained above 90% of field capacity without free water layer on the soil surface. In addition, soil moistures of about 70% and 80% field capacity were set as the control treatments respectively for surface waterlogging (CK1) and subsurface waterlogging (CK2). Logistic regression model was used to simulate the effects of surface waterlogging and subsurface waterlogging on the parameters of grain filling, such as grain filling duration, mean grain filling rate and theoretical maximum 1000-seed weight of summer maize. The results showed that waterlogging stress decreased the duration of grain filling mainly due to decreases in the durations of middle and late grain filling phases. Surface waterlogging for 3 to 5 days decreased grain filling duration by 0.2 to 18.9 days whereas subsurface waterlogging for 5 to 10 days decreased grain filling duration by 2.2 to 7.6 days. Compared with CK, surface waterlogging for 3 days at jointing or tasseling stage resulted in an average increase of 8.2%9.9% in mean grain filling rate. Also for 5 days of surface waterlogging, mean grain filling rate decreased by 10.8%20.9%. Subsurface waterlogging for 5 to 10 days resulted in an average decrease of 0.4%5.2% in mean grain filling rate, where the reduction rate increased with increasing duration of subsurface waterlogging. The theoretical maximum 1000-seed weight and the measured 1000-seed weight increased under 3 days of surface waterlogging, while it reduced under 5 days of surface waterlogging and 5 to 10 days of subsurface waterlogging. The same trends were observed in both the simulated and observed 1000-seed weight. Compared with observed values, simulated results showed that the Logistic regression model overestimated the effect of waterlogging stress on 1000-seed weight.