Consistent evidence from meta-analysis has linked assisted conception by IVF, and particularly intracytoplasmic sperm injection (ICSI), with an increased risk of major birth defects. To compare the risk of major malformations of children born after standard ICSI and after intracytoplasmic injection of morphologically selected spermatozoa (IMSI), a prospective population-based study was conducted from 2005 to 2010. ICSI and IMSI were performed in only one assisted reproduction unit according to its classification of spermatozoa and using fresh semen. Medical data and follow up during 2 years of 1028 infants were collected. Major malformations were identified and classified by an external independent physician. The two groups were similar concerning the parents’ age, treatment, number of oocytes recovered, days of transfer, gestational age and birthweight. However, major malformations were significantly lower with IMSI (6/450, 1.33%) versus ICSI (22/578, 3.80%; adjusted odds ratio 0.35, 95% confidence interval 0.14–0.87, P=0.014), mainly affecting boys (adjusted odds ratio 2.84, 95% confidence interval 1.24–6.53, P=0.009). In conclusion, the significantly decreased risk of major birth defects associated with IMSI remained decreased after multivariate adjustment and highlights the beneficial effect of sperm selection before ICSI.The aim of this study was to compare the birth defects in children born after intracytoplasmic sperm injection (ICSI) and intracytoplasmic morphologically selected sperm injection (IMSI). We conducted a prospective population-based study from 2005 to 2010. Medical data and follow up at 2 years of age for 1028 infants were collected. We show that the major malformations were significantly lower in the IMSI group versus the ICSI group, and they mainly affected the boys. We found less abnormality when abnormal spermatozoa, according to the Cassuto Barak classification, were deselected at high magnification and not injected. In conclusion, the significantly decreased risk of major birth defects associated with IMSI remained decreased after multivariate adjustment and highlights the beneficial effect of spermatozoon selection before ICSI.