In this experimental study, we built a Chinese story reading corpusproduced by 21 Korean learners. Every syllables and vowel sections wereautomatically annotated into Praat textgrids using automaticsegmentation scripts, and two major measurements of rhythmic featureswere calculated from those textgrids. The perceptional foreignaccentedness levels of the speakers were also judged by 3 humanevaluators. The correlations between these rhythmic measurements andforeign accentedness show that the higher the speaker’s proficiency is,the more syllables they read in a second and the lower syllable durarionvariation they make. This suggests that these two rhythmic features canbe effectively used as a parameter in automated speech assessmentsystems for Korean leaners. This study provides implications for L2research since it investigated the contribution of rhythmic features to theperception of foreign accented speech in Chinese, and more importantly,our corpus will be used in the development of automatic speech evaluation systems.