The Chinese Famine of 1959–1961 caused up to 30 million deaths. It varied in intensity across China and affected rural areas disproportionately. Data from the China–U.S. Collaborative Project for Neural Tube Defect Prevention on 31, 449 women (born 1957–1963) and their offspring birth size were recorded in 1993–1996. We used a measure of famine intensity at county level based on the size of famine-born cohorts relative to cohorts preceding and following the famine in a difference-in-difference model that compared offspring birth size of pre-famine (1957–1958; exposed between 0.5 and 4.5 years), famine (1959–1961; prenatal and up to 2.5 years) and post-famine (1962; some exposed in early pregnancy) cohort groups to that of the unexposed 1963 cohort. The model corrected for age and cohort trends and estimated associations between maternal famine exposure and offspring birth size for the average level of famine intensity across counties, and included adjustment for clustering. In rural areas and in pre-famine and famine cohorts, exposure to famine was associated with larger weight (69 g; 95% CI 30, 108), length (0.3 cm; 95% CI −0.0, 0.5) and birth body mass index (0.1 kg/m2; 95% CI 0.0, 0.2). In urban areas, however, exposure to famine was not associated with offspring birth size. Our findings in rural areas suggest that severe and prolonged famine leads to larger newborn size in the offspring of mothers exposed to famine in utero and during the first few years of life; less severe famine in urban areas however, appeared to have no impact. The markedly increased mortality in rural areas may have resulted in the selection of hardier mothers with greater growth potential, which becomes expressed in their offspring.