A laser- induced fluorescence detection set based on liquid core optical fiber was established in this study. Eight edible oils were discriminated by using this detection set combined with chemometrics method. The effect of length of liquid core optical fiber on laser induced fluorescence spectrum was explored, and the differences between the spectra of different edible oils were analyzed. The fluorescence spectra of 320 samples covering 8 types of edible oil were measured in 1 meter liquid core optical fiber. Principal component analysis was used in fluorescence data dimensionality reduction process. Partial least squares discriminant analysis (PLS-DA) method was used to develop the identification model to distinguish edible oil species. The results showed that the oil fluorescence intensity is greatly enhanced when liquid core optical fiber was used. With the increase of liquid core optical fiber length, the peaks of laser induced edible oil fluorescence spectra increased and the fluorescence spectra will produce red shift. The red shift tended to a constant value when the fiber length was more than 80 cm. The fluorescence spectra of different edible oils were quite different, its can be used to distinguish different types of edible oil. Principal component scores chart were get using PC1 and PC2 of edible oils fluorescence data which result in a trend of certain gather of same type of edible oil. The recognition rates of PLS-DA model for the calibration set and prediction set were both 100%. The study shows that the developed device in this study has high accuracy for identifying the edible oil species.