Purpose : Rotavirus is a most common etiologic agent of pediatric gastroenteritis. The standard method to diagnose rotavirus infection was the detection of viral particles in specimens through electron microscopy. But it was complex. Enzyme immunoassay and latex agglutinin are preferred because they are relatively handy, inexpensive and take a short time, in comparison with electron microscopy. However, several reports have shown that the use of ELISA to diagnose rotavirus infection in neonates can result in false positive reactions. The main purpose of this study is to compare ELISA and RT-PCR in the diagnosis of neonatal rotavirus infection. Methods : Data presented in this study were obtained form 123 newborn babies in the nursery of the Fatima Hospital, Masan, Korea, form Jury to December, 1997. We obtained two samples of stool from each of the newborn babies and then performed the Rotazyme test and the RT-PCR. In the Rotazyme test, the results were interpreted according to visual findings. The samples were used for the RT-PCR test after at stock $-30^{circ}C$ to identify rotavirus group A. The result of the two tests were compared. Results : The informations are divided into 73 males and females. Out of the total informations 15 were transferred from other hospitals. Their average gestational age was $38.5{pm}1.6$ weeks. The average birth weight was $3134.8{pm}539gm$. In the Rotazyme test, 75 samples turned out to be positive. Out of them, 55 samples(75.3%) were positive and 18 samples(24.7%) were negative in the RT-PCR. On the other hand, in the Rotazyme test, 50 samples turned out be negative. Out of them, 27 samples(54%) were positive and 23 samples(46%) were negative in the RT-PCR. Conclusion : Rotavirus infection in uncommon in neonates. The diagnosis based on visual findings using Rotazyme test has a disadvantage in the sense that it can result in false positive reactions and false negative reactions in the diagnosis of neonatal rotavirus infection.