Future change in partnerships and fertility are not easy to forecast. However, the fertility of the youngest cohorts will depend on those behaviours. The way young people start a partnership has changed a lot during the past three decades. Many couples are now unmarried, union disruptions and step-families are more frequent, young people leave school later and the age of motherhood has increased. Microsimulation can provide a measure of the change in future completed fertility, which helps to analyse the influence of current behaviours on future change in family structures. If behaviours remain the same as the ones observed from 1995 to 1996, completed fertility may decrease to less than 2 children per woman born around 1970 and remain constant about 1.9 children per woman for women born after 1975. This decrease stems from a postponement in the age of motherhood and an increase in union disruptions. Fertility at older ages as well as the desire to have at least one common biological child in step-family do not offset the negative effects of delaying births and living longer without a partner before the age of 45. Timing in unions and disruptions seems to play an important role in fertility. Women who live only a short time without a partner after a disruption experience have practically the same number of children as those who are still living with their first coresident partner. These simulations remain of course prospective and rely on several assumptions which may not prove adequate in the future. In particular, they assume that future behaviours will remain identical to the ones currently observed. Although estimated behaviours provide results consistent with available empirical work, it does not mean that some household behaviours will not change in the long-run.