BACKGROUND: Concerns about possible adverse outcomes for children conceived using ICSI were highlighted in 1998 when 1-year-old ICSI children were found to be at increased risk (relative risk = 9.2) of delayed mental development compared with children conceived naturally or using IVF. As the findings were biologically plausible, it was considered important to reassess child development when a more accurate measure of long-term cognitive ability could be obtained. METHODS: The mental development of 97 ICSI, 80 IVF and 110 naturally conceived (NC) children at 5 years of age was assessed using intelligence quotients (IQ) obtained from the Wechsler Preschool and Primary Scales of Intelligence. RESULTS: The mean full-scale IQ was 110 ± 18 for ICSI, 111 ± 13 for IVF and 114 ± 13 for NC children (P = 0.21, non-significant). ICSI children were not at increased risk for delayed (full-scale IQ <85) cognitive development (ICSI 5.2%, IVF 2.5%, NC 0.9%; P = 0.18, non-significant). The only significant independent predictor of below-average full-scale IQ on multivariate analysis was lower maternal education level. CONCLUSIONS: These findings suggest that the genetic influence of parental cognitive ability is more important than the mode of conception in determining the long-term intellectual ability of children conceived using ICSI.