Early marriage and fertility are major social determinants of health and wellbeing. Rapid shifts in the past three decades, including a rise in sexual activity in unmarried adolescents, a large population of young migrant workers, and a high proportion of males relative to females, have the potential to alter patterns of reproductive health in Chinese adolescents and young women. We aimed to establish long-term trends of marriage and fertility for girls and women aged 15-24 years in China. We did a longitudinal study in which we extracted aggregated data for marriage and childbearing status for Chinese girls and women aged 15-24 years from the Chinese National Population Census (in 1990, 2000, and 2010) and the Chinese 1% National Population Sample Survey (in 1995, 2005, and 2015). The census included all individuals with Chinese nationality who resided in China when the survey was done. For the 1% sample survey, communities or villages were randomly selected and all residents with Chinese nationality in the selected communities or villages were included. In all censuses and sample surveys, forms that included information on basic demographic characteristics, education, marriage, and fertility were completed and verified by the census enumerators at the household residence, based on responses provided by the householder or another adult in the household. We calculated the ever-married rate and age-specific fertility rate (ASFR) for all included individuals. We built multivariate random-effects generalised least squares regression models on panel data to test whether marriage or fertility rate was associated with education level, sex ratio, and the proportion of the population who are an ethnic minority in a province. The ever-married rate for those aged 15-19 years decreased from 4·7% in 1990 to 1·2% (95% CI 1·2-1·3) in 2000, but rebounded to 2·4% (2·4-2·5) in 2015. The ASFR for this age group decreased from 22·0 births per 1000 individuals in 1990 to 6·0 (5·9-6·0) births per 1000 in 2000, and rebounded to 9·2 (8·9-9·4) births per 1000 in 2015. The rebound was found in most provinces. In women aged 20-24 years, the ever-married rate generally declined from 58·6% in 1990 to 25·5% (95% CI 25·4-25·6) in 2015, and the ASFR decreased from 198·8 births per 1000 in 1990 to 55·0 (54·5-55·5) births per 1000 in 2015. In 2015, the ever-married rate and ASFR for girls and women in rural areas aged 15-19 years were three-times higher than those of their urban counterparts (3·8% [95% CI 3·7-3·9] vs 1·1% [1·1-1·1] for the ever-married rate and 15·4 [14·9-15·9] vs 4·1 [3·9-4·3] births per 1000 for the ASFR). There were large disparities in ever-married rate across individuals of different education levels, with increases in the ever-married rate of 15-19-year-olds not attending senior high school between 2000 and 2010. Those aged 15-19 years were more likely to be married or give birth in the western provinces. Education held a protective association against adolescent childbearing, whereas a high ratio of males to females and a high proportion of ethnic minorities were associated with greater risk. Although China's total fertility rate remains far less than replacement, after a period of steady decline, there has been a rebound in adolescent marriage and childbearing in the past decade. A range of adolescent-targeted strategies will be needed across provinces, including scaling up comprehensive sex education, ensuring that girls continue in school, and providing adequate reproductive health services, which specifically meet the needs for modern contraception in sexually active unmarried girls. Humanities and Social Sciences Planning Fund Project, Sports and Health Special Project of Education and Scientific Research, and China Scholarship Council.