The aim of this study was to compare sagittal morphometric features between Liza haematocheilus stocks in the China coasts at the Ying Kou, Dong Ying, Qing Dao, Wen Zhou and Guang Zhou regions and test a complementary method to quantify the variation of sagittal shapes to discriminate L. haematocheilus stocks. The sagitta variation of five L. haematocheilus stocks was examined using nine shape indices (Roundness, Circularity, Form-factor, Rectangularity, Ellipticity, Radius ratio, Feret ratio, Aspect ratio and Surface density). Multiple comparisons on shape indices showed that there was no significant difference among the Ying Kou and Dong Ying stocks, possibly due to their similar living habitat in the Bohai Sea. As the distribution zones of these five stocks are much overlapped, principal component analysis for otolith shape is not possible to distinguish L. haematocheilus stocks. Based on the Fourier coefficient, Fisher’s discriminant analysis accurately classified 76.2% specimens into correct stocks whereas the discriminant function of shape indices only correctly identified 62.8% specimens into right stocks. As the sagittae in L. haematocheilus had an irregular round shape with many sharp notches, this study indicates that Fourier analysis is more suitable than shape indices to discriminate sagittae with irregular shapes. Our results demonstrate that as long as substantial intraspecific variations exist in sagittae shapes, geometric morphometrics for otolith shapes could be used as a complementary tool along with body morphology to distinguish L. haematocheilus stocks.