Objective To evaluate the effect of music therapy for childbirth. Methods Such databases as The Cochrane Library, PubMed, EMbase, EBSCO host, SpringerLINK Online Journals, CBM and WanFang Data were searched from January of 2000 to December of 2010 to collect randomized controlled trials (RCTs) of music therapy for childbirth. The quality of RCTs was appraised and the data were extracted. Meta-analyses were conducted with RevMan5.02 software for the standarded RCTs. Results A total of nine RCTs were included. Five RCTs indicated the music therapy could alleviate the labor pain; five RCTs indicated the music therapy could reduce the event risk of cesarean section due to the failure of transvaginal trial labor; three RCTs indicated the music therapy could shorten the first stage of labor; two RCTs indicated the music therapy could stabilize the systolic pressure and heart rate when complete cervical dilation was done, and three RCTs indicated the music therapy could relieve anxiety. In addition, music therapy had no influence on neonate Apgar’s score; and the result of meta-analyses on postpartum hemorrhage was not reliable through sensitivity analyses. Conclusion The music therapy applied during childbirth can relieve the labor pain and anxiety, stabilize the heart rate and systolic pressure when complete cervical dilation is done, reduce the event risk of cesarean section due to the failure of transvaginal trial labor, shorten the first stage of labor, and is beneficial to the mind and body of parturient.