Preliminary evidence indicates that natural disasters are associated with an increased risk for schizophrenia. With few longitudinal studies on earthquakes, this retrospective cohort study examined exposure to the 1976 Tangshan earthquake and the subsequent risk of schizophrenia. Population counts and visits to all nine psychiatric hospitals in Tangshan city were collected. We created three cohort groups by earthquake exposure: infant (August 1972 to July 1976 births), fetal (August 1976 to May 1977 births), and unexposed (June 1977 to May 1981 births). The cumulative incidence of schizophrenia in each cohort was calculated by dividing the number of schizophrenia patients by total births in the corresponding period. Altogether, 6424 schizophrenia patients were identified, with 2786 in the infant group, 663 in the fetal group, and 2975 in the unexposed group. The crude cumulative incidence of schizophrenia in the infant, fetal and unexposed groups were 7.64 (95% confidence interval [CI] = 7.36–7.92), 9.07 (95% CI = 8.38–9.76), and 7.40 (95% CI = 7.13–7.66) per thousand population respectively. Adjusted for mortality, the corresponding figures were 7.73 (95% CI = 7.44–8.01), 9.30 (95% CI = 8.60–10.01) and 7.44 (95% CI = 7.18–7.71) per thousand population respectively. The mortality-adjusted risk ratio (aRR) was 1.25 (95% CI = 1.15–1.36) between fetal and unexposed groups (χ2 = 27.31, P < 0.001). Males exposed as infants did not differ from the unexposed in cumulative schizophrenia incidence. People with fetal exposure to the 1976 earthquake had 25% higher risk of developing schizophrenia compared to unexposed counterparts.