Background: Because pregnancy is a predictable life event and perinatally-occurring OCD is highly prevalent, the perinatal period is an excellent time to investigate OCD development. The cognitive-behavioural (CBT) model of OCD development posits that negative beliefs about unwanted, intrusive thoughts (UITs) increase the likelihood of OCD development. Previously published studies provide support for this model but key knowledge gaps remain.Objectives: To assess prenatal beliefs about thoughts as predictors of postpartum OCS and test the bidirectional relationship between beliefs about thoughts and OCS.Methods: Participants were an unselected sample of pregnant people in British Columbia (N = 760). Assessments occurred in late pregnancy and twice postpartum and included self-report measures of OCD (DOCS) and obsessive beliefs (OBQ-44). Random intercept cross-lagged panel models tested bidirectional associations between beliefs about thoughts and OCS over the perinatal period, controlling for prior symptoms and covariates.Results: Prenatal OBQ-44 scores predicted both early and later postpartum DOCS scores. OBQ-44 scores better predicted later postpartum OCS, compared with early postpartum OCS. The relationship between beliefs about thoughts and OCS was found to be largely independent of depressed mood. Prenatal OBQ-44 scores better predicted both early and later postpartum DOCS scores than did prenatal DOCS scores predicting postpartum OBQ-44. Findings related to DOCS subscales partially supported study hypotheses.Conclusions: Findings provide further support for the CBT model of OCD and suggest that negative beliefs about thoughts (rather than OCS) represent an important mechanism in postpartum OCS severity. Study findings have implications for prevention and treatment.