Abstract Purpose This study aimed to describe the implementation of a new retrospective Belgian national cohort of pregnant women, the Belgian Medication Exposure during Pregnancy (BeMeP). Methods We linked the national dispensing data to birth and death certificates and hospital stay data for a 7‐year period between 2010 and 2016 for the first time in Belgium. We presented the characteristics of pregnancy events associated with the mothers enrolled in the linkage study. Next, we constructed a cohort of pregnancies and compared some characteristics computed using the BeMeP database with the national statistics. Finally, we described the use of medications during pregnancy based on the first level of the Anatomical Therapeutic Chemical (ATC) classification. Results We included 630 457 pregnant women with 900 159 pregnancy‐related events (843 780 livebirths, 1937 stillbirths, 6402 ectopic events, and 47 905 abortions) linked to medication exposure information. Overall, 96.3% of live births and 83.5% of stillbirths (national statistics as reference) were captured from the BeMeP. During pregnancy, excluding the week of birth, 78.9% of live birth pregnancies and 79.6% of stillbirth pregnancies were exposed to at least one medication. The most frequently dispensed medications were anti‐infectives (ATC code J = 50.2%) for live births and for stillbirths (44.0%). Conclusion We linked information on pregnancies, all reimbursed medications dispensed by community pharmacists, all medications dispensed during hospitalization, sociodemographic status, and infant health to create the BeMeP database. The database represents a valuable potential resource for studying exposure‐outcome associations for medication use during pregnancy. ; info:eu-repo/semantics/published