In this study, we explored the predictive role of family interactions and family representations in mothers and fathers during pregnancy for postnatal motherfatherinfant interactions during the first 2 years after birth. Families (N = 42) were seen at the fifth month of pregnancy and at 3 and 18 months after birth. During pregnancy, parents were asked to play with their baby at the first meeting by using a doll in accordance with the procedure of the prenatal Lausanne Trilogue Play (LTP; A. Corboz-Warnery & E. Fivaz-Depeursinge, 2001; E. Fivaz-Depeursinge, F. Frascarolo-Moutinot, & A. Corboz-Warnery, 2010). Family representations were assessed by administering the Family System Test (T. Gehring, 1998). Marital satisfaction and the history of the couple were assessed through self-reported questionnaires. At 3 and 18 months, family interactions were assessed in the postnatal LTP. Infant temperament was assessed through parent reports. Results show that (a) prenatal interactions and child temperament are the most important predictors of family interactions and (b) paternal representations are predictive of family interactions at 3 months. These results show that observational assessment of nascent family interactions is possible during pregnancy, which would allow early screening of family maladjustment. The findings also highlight the necessity of taking into account paternal representations as a significant variable in the development of family interactions.