Abstract: Objective To understand the trend of underreporting of death causes in disease surveillance areas, Hubei province, from 2009 to 2014. Methods In 2012, according to the average population and social economic level of townships (communities) in disease surveillance areas, 1 township (community) was selected from 3 townships (communities) with similar crude mortality rates in each disease surveillance area, a household survey was conducted in all the families in the township (community) selected. In 2015, according to the same standard, 9 villages (neighborhoods) were selected from 3 townships randomly selected in each disease surveillance area, a household survey was conducted in all the families in 9 villages (neighborhoods). So the underreporting of death causes during 2009-2011 and during 2012-2014 were surveyed respectively. Results Compared with the period of 2009-2011, the underreporting rate of death causes during 2012-2014 decreased with year and the annual decrease had statistical differences(2=38.09, P=0.00). From 2009 to 2014 the underreporting rates of death causes decreased in both urban area and rural area in Hubei(2=37.92, P=0.00), but the underreporting rate in urban females and the gender specific underreporting rates in rural area had no obvious decrease (2=2.82, P=0.09; 2=2.37, P=0.12; 2=1.48, P=0.22).The underreporting rate of death causes in people aged 5 years decreased with year, the annual decrease had statistical differences(2=38.68, P=0.00), but the decrease in young children or infants aged 5 years was not obvious. The underreporting rate for injury deaths was higher than that for disease caused deaths(2=5.61, P=0.02). Conclusion The overall underreporting rate of death causes decreased with year in Hubei during 2009-2014. The underreporting rates in age group 5 years and in rural population were still high. The underreporting rate for injury deaths was higher than that for disease caused deaths. It is necessary to take appropriate measures to reduce the death cause underreporting and improve the completeness of death cause data.