Introduction: Measuring early-life psychosocial stress is complicated by methodological challenges. This paper compares three survey instruments for the assessment of life in pregnancy/postpartum and investigates the effects of the timing of early-life stress for emotional/behavioral difficulties (EBD) of offspring during mid/late childhood and adolescence. Methods: Observational data were obtained from the European Longitudinal Cohort Study of Pregnancy and Childhood (ELSPAC-CZ), which included 4811 pregnancies in two Czech metropolitan areas. We used data collected between 1991 and 2010 at 20 weeks of pregnancy (T1), after delivery (T2), at 6 months postpartum (T3), and at child’s age of 7 years (T4), 11 years (T5), 15 years (T6), and 18 years (T7). Life stress was assessed with (1) the Edinburgh Postnatal Depression Scale (EPDS), (2) a stressful life events (SLE) count based on 42-item inventory, and (3) the SLE measure weighted by perceived stressfulness (PS). Each stress measure was administered at T1, T2, and T3. Child’s EBD were assessed with the Strengths and Difficulties Questionnaire at T4, T5, T6, and T7. Results: Each stress measure independently predicted long-term EBD. The best data fit was obtained in a model combining EPDS and SLE. Effect sizes for SLEs decreased between the first half of pregnancy and postpartum, while the effect of EPDS increased. Discussion: SLE-based methods capture an aspect of perinatal stress not adequately assessed by EPDS. Combination of psychological distress measures and SLE-based measures is optimal in predicting EBD of the child. Stress measures based on SLE are suitable for early pregnancy, while self-reports of depressive symptoms may perform better in postpartum.