China is highly susceptible to flood disasters and subjected to great damage every year. Furthermore, the flood frequency has exhibited an increasing trend in recent years. Most flood events, including flash floods and river flood, are induced by rainfall. This study investigates annual variations of rainfall occurrence over China during the period from 2000 to 2015 at the national and regional scale using daily rainfall data from the Tropical Rainfall Measuring Mission. The Mann-Kendall test is performed for trend detection, and statistical data of flood damage published by China’s government, including destroyed crop area, damaged buildings, direct economic loss, percentage of GDP (gross domestic product), and death toll are correlatively analysed with rainfall occurrences. The results show that storm rain events show the greatest variation among three rainfall types (moderate rain, heavy rain and storm rain). The variation coefficients of rainfall over Northeast China, North China, and Northwest China are the highest, whereas that for Southwest China is the smallest. Moderate rain, heavy rain over Central China, and moderate rain over Southwest China exhibits decreasing trends, whereas the remaining exhibit increasing trends. The correlation between the rainfall occurrences and these flood damage indices at the national scale shows that only direct economic loss has a strong positive correlation with rainfall occurrences, and the other indices have weaker correlations. The correlation is strong in three north regions, except for death toll in Northwest China. In contrast, the correlation between flood damage and rainfall is weak in East China, Central China, Southwest China, and South China. Overall, death toll is strongly correlated with the number of damaged buildings, implying that flood fatalities in China are likely associated with building collapse, and are dominated by specific extreme events. This study can provide a scientific reference for flood management in China.