Psychological tests based on visual information processing have shown to be promising in predicting neurodevelopmental outcome in infants at risk. In the present study we prospectively investigated the early development in a group of 20 high-risk preterm infants by means of i) the Fagan Test of Infant Intelligence at 7, 9, and 12 months postterm and ii) a detailed battery for the early assessment of visual functions at 6 and 10 months postterm. The results were then correlated to the Griffiths development scales at two years. At around 7 months no correlation was found in our infants between the Fagan test and neurodevelopmental outcome, possibly as a consequence of the influence of abnormal oculomotor behaviour. At around 9 months most of the visual abnormalities were no more present and the Fagan test was significantly correlated with the outcome. At 12 months postterm a decline of the predictive value of the FTII was observed. In conclusion, nine months postterm age appears to be the best age for the early assessment of neurodevelopmental outcome in high-risk preterm infants, as the maturation of the attentional and visual systems allows a more reliable evaluation.