Objective To investigate whether early intervention on premature very low birth weight infants can reduce the incidence of cerebral palsy. Methods The sample data in this study included 857 premature infants with gestational age less than 32 weeks and/or birth weight lower than 1 500 g from 28 hospitals.The children with congenital malformations and hereditary diseases were excluded.All children were divided into two groups: intervention group (452) in which their parents actively received early intervention guidance and control group (405) in which all children were born 9 to 15 months prior the study and were only given conventional parenting.A comprehensive early intervention program was performed to children in intervention group after discharge from hospitals,including the training of motor,cognitive,language,emotional and communication abilities.Among them,Motor training included massage,passive exercise and active exercise training.Children were then taken neuromotor examination once a month.Children with abnormal posture received intensive training by professional rehabilitation personnel; children in control group only received the same feeding,care and infant health guidance with children in the intervention group.Subjects were then taken neuromotor examination on the corrected age of 9 and 12 months respectively.All children were followed up to 2 years old. Results Children in two groups were without difference in gender,gestational age,birth weight and other general aspects (P>0.05).Perinatal period comparison:There were 118 (26.33%) and 75 (18.43%) children were found with intracranial hemorrhage in the intervention and control group respectively.The occurrence of intracranial hemorrhage in the intervention group was significantly higher (P<0.05).The comparisons of the incidence ratios of intrauterine distress,postnatal asphyxia,hypoxic-ischemic encephalopathy,hyperbilirubinemia,respiratory distress syndrome and neonatal infection were statistically insignificant (all P>0.05).The diagnoses of cerebral palsy were given only when children were older than 1 in both groups.The incidence of cerebral palsy was 22.1 ‰(10/452) in the intervention group,and 61.7 ‰(25/405) in control group,the incidence ratio in the control group was 2.79 times higher,and the difference was statistically significant (P<0.01). Conclusion Early intervention can effectively reduce incidence of cerebral palsy in premature children and very low birth weight children after discharge.