The reference sound pressure level (RSPL) for speech audiometry is established based on the speech recognition threshold (SRT) measured by standardized words and procedures (ISO 8253-3-2012). The aim of this study was to obtain the basic data for standardization of the reference sound pressure level (RSPL) for Korean speech audiometry which was defined as the reference speech recognition threshold level (RSRTL) regarding O dB HL at the audiometer. Subjects consisted of 230 adults with normal hearing sensitivity aged between 18 and 25 years from 5 universities. Puretone thresholds were obtained in 2 dB steps at each octave band. The newly developed Korean bisyllabic words were used for SRT test. SRTs were measured by the descending method described in ISO 8253-3 only at the better ear based on the PTA or right ear if PTAs were equal using 36 bisyllabic words which consist of 3 sets and each set has 12 test items. Retest of SRT was performed for the same subject about an hour after the first test. The RSPL was determined by the median of all subjects SRTs. Correlation analysis, paired t-test and confidence interval (CI) were used for test-retest reliability. Results suggest that the test-retest reliability of SRT be high and the RSPL for Korean speech aiudiometry be 23 dB SPL.