Many female survivors of adolescent and young adult cancers (AYA survivors) have shortened reproductive lifespans. However, the timing and duration of ovarian function after cancer treatment are largely unknown.To model the trajectory of ovarian function over two decades following cancer treatment and evaluate how trajectories vary by treatment gonadotoxicity and age.In a prospective cohort, AYA survivors ages 18-39 at variable times since cancer treatment completion provided dried blood spots (DBS) every 6 months for up to 18 months. AMH levels were measured using the Ansh DBS AMH ELISA assay. The mean AMH trajectory was modeled for the entire cohort and separately by treatment gonadotoxicity and age using functional principal components analysis.763 participants, mean (SD) enrollment age 33.3 (4.7) and age at cancer diagnosis 25.9 (5.7) years, contributed 1905 DBS samples. The most common cancers were breast (26.9%), lymphoma (24.8%) and thyroid (18.0%). AMH trajectories differed among survivors by treatment gonadotoxicity (low, moderate or high) (p<0.001). Following low or moderately gonadotoxic treatments, AMH levels increased over 2-3 years and plateaued over 10-15 years before declining. In contrast, following highly gonadotoxic treatment, AMH levels were lower overall and declined shortly after peak at 2-3 years. Younger age at treatment was associated with higher trajectories, but a protective effect of younger age was not observed with in survivors exposed to highly gonadotoxic treatments (pinteraction <0.001).In this large AYA survivor cohort, timing and duration of ovarian function strongly depended on treatment gonadotoxicity and age at treatment. The findings provide novel, more precise information to guide reproductive decision-making.