【Objective】 Investigate the applying value of prechtl's qualitative general movements assessment(GMs) and brainstem auditory evoked potential(BAEP) in predicting of cerebral palsy. 【Methods】 According to the inclusion and exclusion criterias, infants who took part in our follow-up clinic after discharged from our neonatal intensive care unit were confirmed as the participants. Each infant had GMs and BAEP examination. The follow up lasted for at least one year to identify neurological outcome. The diagnosis of cerebral palsy was made according to the international cerebral palsy diagnostic definition published in 2005. Use SPSS 13.0 to analysis two examination's positive predictive values, negative predictive values, specificity and sensitivity. 【Results】 Twenty-seven high risk infants were followed up, eight of them were diagnosed as cerebral palsy. The predictive validity of GMs was as following: sensitivity 87.5%, specificity 84.2%, positive predictive value 70.0% and negative predictive value 94.1%. The predictive validity of BAEP was as following: sensitivity75.0%, specificity36.8%, positive predictive value 33.3% and negative predictive value 77.7%. The abnormalities of BAEP mainly were elevating of auditory threshold and prolonged latency. 【Conclusions】 GMs assessment can be used in high risk infants within 4 to 5 months to give a reliable and valid prediction for later neurological outcome, especially for cerebral palsy. Comprehensive application of GMs and BAEP can relatively overall impression of infants' function status.