OBJECTIVE: To develop a method for estimating the cross-sectional number of pregnant women, distributed by gestational age for cost-effectiveness and budget impact analysis. METHODS:   We obtained data on the monthly number of deliveries at gestational ages 20-42 weeks for years 2003 through 2006 (the most recent available) from publicly accessible birth certificate data from the US National Center for Health Statistics (NCHS).  Estimates of the number of deliveries at each gestational age each week were derived based on the number of days in each month, under the assumption that deliveries were randomly distributed throughout the month.  Cohorts of women who became pregnant during the same calendar week were constructed using the following formula: Deliveries at 20 weeks (calendar week X) + Deliveries at 21 weeks (calendar week X+1) + ….Deliveries at 42 weeks (calendar week X+22). This procedure was repeated for each week from January 2003 through July 2006.     Estimates of the number of pregnancies between 6 and 20 weeks were generated by applying gestational age-specific loss rates from a large prospective cohort study of early pregnancy. RESULTS:  Over the 3.5 year period, there were approximately 2.5 million pregnant women ranging from 6 to 42 weeks gestational age in the US; the mean weekly number increased from 2.5 to 2.6 between 2003 and 2006.  There was also substantial seasonal variation in singleton pregnancies, with the highest number of pregnant women observed in March and April of each year, but no seasonal variation in multiple gestations.   CONCLUSIONS:  Routinely collected data on births allows estimation of the number of pregnant women at any point in time.   Such data, in conjunction with estimates of probabilities of delivery and pregnancy and neonatal complications, can be useful for cost-effectiveness and budget impact analyses of interventions during pregnancy.