Human is the key to social and economic development, and plays a decisive role in rural evolution. However, influenced by urban-rural dual structure, there are large population outflow in rural China, which makes the strategy of rural revitalization lack of human capital support. Therefore, to scientifically promote the implementation of rural revitalization in the new era, it is necessary to study the rural population outflow. Using datasets of 2000 and 2010 population census of the People's Republic of China by county, this research analyzed the evolution and spatial pattern of rural population outflow in China. Then it explored the mechanism of regional differentiation of outflow rate of rural population (ORRP), which was reflected by the ratio of rural hukou-registered population and rural residential population. Results showed that the development of rural population in China from 1978 to present could be divided into three stages i.e., slow development stage (1978-1995), massive outflow stage (1995-2014) and citizenship of agricultural transfer population (since 2014). From 2000 to 2010, ORRP at county level in China generally showed an upward trend, the centralization of rural hukou-registered and residential population declined obviously. From the perspective of spatial pattern, they all showed significant spatial clustering and positive spatial correlation. In 2000, there were 76.06% counties with ORRP ranging from 100% to 150%, and only 14.01% of the counties' ORRP were less than 100%, which were mainly located in Qinghai-Tibet Plateau, Xinjiang and the border areas of northeast China. By 2010, the proportion of counties with ORRP less than 100% fell to 5.20%, concentrating in Xinjiang, and there were also a few counties in Qinghai-Tibet Plateau and Northeast China. Based on pre-selected influencing factors, estimated results of spatial error model (SEM) indicated that elevation, per capita GDP, urban-rural income ratio, proportion of primary industry, urbanization rate, road density, educational attainment and per capita investment in fixed assets were positively correlated with ORRP at county level, while per capita farmland, per capita net income of rural households and ratio of rural employees showed a negative correlation.