Background: Chronic obstructive pulmonary disease (COPD) is caused by interactions between many factors across the life course, including genetics. A proportion of COPD may be due to reduced lung growth in childhood. We hypothesized that a polygenic risk score (PRS) for COPD is associated with lower lung function already in childhood and up to adulthood. Methods: A weighted PRS was calculated based on the 82 association signals (p ≤ 5 × 10−8) revealed by the largest GWAS of airflow limitation (defined as COPD) to date. This PRS was tested in association with lung function measures (FEV1, FVC, and FEV1/FVC) in subjects aged 4–50 years from 16 independent cohorts participating in the Chronic Airway Diseases Early Stratification (CADSET) Clinical Research Collaboration. Age-stratified meta-analyses were conducted combining the results from each cohort (n = 45,406). These findings were validated in subjects >50 years old. Findings: We found significant associations between the PRS for airflow limitation and: (1) lower pre-bronchodilator FEV1/FVC from school age (7–10 years; β: −0.13 z-scores per one PRS z-score increase [–0.15, −0.11], q-value = 7.04 × 10−53) to adulthood (41–50 years; β: −0.16 [–0.19, −0.13], q-value = 1.31 × 10−24); and (2) lower FEV1 (from school age: 7–10 years; β: −0.07 [–0.09, −0.05], q-value = 1.65 × 10−9, to adulthood: 41–50 years; β: −0.17 [–0.20, −0.13], q-value = 4.48 x 10−20). No effect modification by smoking, sex, or a diagnosis of asthma was observed. Interpretation: We provide evidence that a higher genetic risk for COPD is linked to lower lung function from childhood onwards. Funding: This study was supported by CADSET, a Clinical Research Collaboration of the European Respiratory Society.