The purpose of this study is to investigate the difference in the pronunciation of the Chinese velar fricative /h/[x] between a native speaker of Chinese and a Korean advanced learner of Chinese. It is believed that Korean learners of Chinese assimilate the Chinese velar fricative [x] as the Korean glottal fricative [h], whose acoustic characteristics are influenced by the subsequent vowels. Accordingly, it can be assumed that the pronunciation of the Chinese [x] by Koreans learners is not always different from that by a native speaker. To assess this allophonic difference, an experiment was conducted involving the production of the Chinese [x] with several different subsequent vowels. The sound data recorded from the experiment was analyzed according to four acoustic phonetical cues based on the “source-filter theory” of Fant(1960) and Flanagan(2013), that is, the frication duration, the center of gravity in the spectrum, the skewness of the spectrum and formant values. First, it was found that the average frication duration of the Chinese [x] by Chinese native speakers is significantly longer than that by Korean learners of Chinese in general (p<.05). However, when it is followed by the low vowels [a, A, α], no significant difference was found owing to pharyngealization in Chinese. Second, the F1 and F2 values of the Chinese /h/[x] by Korean learners were generally lower than those by Chinese native speakers, indicating that the place of articulation of the Chinese [x] by Korean learners is at a lower point of the tube compared to that of a Chinese native speaker. Third, there were no significant differences in the values of the center of gravity or skewness in the spectrum only when [x] is followed by a back vowel, such as [u, υ, γ]. This was found to occur due to the phonological transfer from Korean, which changes the place of the articulation of /h/ from the glottis to the velar in front of back vowels. To conclude, the production of the Chinese velar fricative by Korean learners shows partial differences depending on the succeeding vowels. This finding can shed new light on Chinese pronunciation education in Korea.