Purpose : The aim of the study is to compare the effect of cardiopulmonary resuscitation (CPR) with voice and CPR without voice by one rescuer. Methods : Subjects were 26 students in C University who had basic life support certificate for Healthcare Provider. They performed 30:2 CPR for 6 minutes by two groups of CPR with voice and CPR without voice by one rescuer from August 14 to 16, 2012. They performed CPR with Resusci Anne SkillReporter$^{TM}$ and Laerdal PC SkillReporting System Ver. 2.4.1(Laerdal Medical, Norway and recored voice using TES-1350A(TES Electrical Electronic Corp, Taiwan). Between each experiment, 1 day of rest was given, providing enough time to recover from the fatigue of CPR. Results : The depth, rate of chest compression, and ventilation volume were not affected by a voice (p >.05), and the ratio of chest compression to ventilation kept 30:2, when the subject made a sound (p <.05). Conclusion : Making voice during CPR was associated with an accurate ratio of 30:2 and the reduction in hands off time.