Abstract: Based on 9 morphological characters, morphological variations of four populations (Guangxi, GX; Leizhou, LZ; Yuenan,YN; Sanya, SY) of pearl oysters Pinctada fucata were investigated using multivariate morphometrics. As the result of principal component analysis (PCA), three principal components were constructed by factor loadings, in which the first principal component (PC) was affected by 7 characters, PC2 and PC3 each by 1 character. The contribution ratios of three PCs were 55.577%, 14.475% and 10.545%,respectively. The results of PCA and cluster analysis revealed that the morphological characters of LZ and YN populations were similar and much more different from SY population. The discriminant functions of four P.fucata populations were established, and the discriminant accuracy was 50%, 58%, 60%, 60% for GX, LZ, YN and SY, respectively, with the average discriminant accuracy being 57%. The morphological variation of P.fucata was a result of the joint effect of genetic components and environment, so it could be a reference for broodstock selection in genetic improvement.