Abstract: Characteristics of droughts and floods in China during the summer (May-August) of 2016 and 1998 were compared in great detail. The associated atmospheric circulation and external-forcing factors were also investigated and compared for these two years. (1) The precipitation was mostly above normal in China in the summer of 2016, with two main rainfall belts located in the Yangtze River valley (YRV) and North China. Compared with the summer of 1998, a similar rainfall belt occurred in the YRV with the precipitation over 100% above normal. However, the seasonal processes of Meiyu were different. The typical "second Meiyu" occurred in 1998, whereas dry condition dominated the YRV after the end of Meiyu in 2016. (2) During May-July 2016, the Ural high was weaker than normal, but it was stronger than normal in 1998. This difference was resulted from the totally different distribution of sea surface temperature anomaly (SSTA) over the North Atlantic Ocean during the preceding winter and spring of the two years. (3) However, the tropical and subtropical circulations were much similar in May-July of 2016 and 1998. Circulations in both years were characterized by stronger than normal and more westward-extending west Pacific subtropical high (WPSH), weaker than normal East Asian summer monsoon (EASM), and anomalous convergence of moisture flux in the middle and lower reaches of the YRV. These similar circulation anomalies were attributed to the similar tropical SSTA pattern in the preceding seasons, i.e. the super El Niño and strong warming in the tropical Indian Ocean. (4) Significant differences in the circulation pattern were observed in August between the two years. The WPSH broke in August 2016, with the western part combining with the continental high and persistently dominating eastern China. The EASM suddenly became stronger, and dry condition prevailed in the YRV. On the contrary, the EASM became weaker in August 1998 and the "second Meiyu" took place in the YRV. The Madden-Julian Oscillation (MJO) activity became extremely active in August 2016 and stayed in the western Pacific for 25 days. It triggered frequent tropical cyclone activities and further influenced the significant turning of the tropical and subtropical circulation in August 2016. However, MJO was active in the tropical Indian Ocean in August 1998, which was favorable for the maintenance of a strong WPSH. Except for the above ocean factors and atmospheric circulation anomalies, the thermal effect of the snow cover over Qinghai-Tibet Plateau from the preceding winter to spring in 2016 was much weaker than that in 1998. This may explain the relatively stronger EASM and more precipitation in North China in 2016 than in 1998.