Paralichthys olivaceus is one of the most important aquaculture species in China. However, intensive breedingand many generations of inbreeding have resulted in poor disease resistance and slow growth. Therefore, it is necessaryto improve the quality and productivity of P. olivaceus through selective breeding. To study the relationship betweenmorphometric traits and body weight of P. olivaceus at different growth stages, 83 fish were selected randomly (8months old, n = 52 and 14 months old, n = 31) to measure 22 morphometric traits and body weights. The data were statisticallyprocessed using correlation, path, and stepwise regression analyses, and multivariate regression equations werecalculated. The traits were included in the equation as independent variables, and body weight was the dependent variablein six curve-fitting models to select the optimal model. The results showed that: (1)the correlation coefficients betweenbody weight and each trait varied and differed at the two developmental stages. (2)The path coefficients of X18,X4, and X11 for 8-month body weight were highly significant (P<0.01). The path coefficients of X18, X14, and X9 for14-month body weight were also significant (P<0.05). The multivariate regression equation was Y = −119.541 + 7.191X18 + 10.135 X4 + 7.197 X11 for 8 months, and Y = −484.931 + 31.959 X18 + 81.928 X14 − 17.889 X9 for 14 months. Thisresult suggests that the vital morphometric traits that affected body weight at the two growth stages were different.(3) All optimal models at 8 months were linear, including Y = −117.866 + 15.724 X18, Y = −94.579 + 24.763 X4, and Y =−100.602 + 33.184 X11. However, all optimal models at 14 months were power models, including Y = 0.036 X183.063, Y =0.095 X142.507, and Y = 62.249 X91.584. These results indicate that the growth rhythms between the two growth stageswere different. The R2 values for the multivariate regression equations at 8 and 14 months were 0.965 and 0.984, respectively,indicating that the equations can be used to predict P. olivaceus body weight. The R2 values for the X11linear model at 8 months and the X9 and X14 power models at 14 months were all < 0.85, indicating that X11, X9, and X14did not explain body weight as a single independent variable. The X18 path coefficient appeared in the multivariate regressionequations at both growth stages, indicating that X18 is a predominant morphometric trait affecting body weightand can be regarded as an assisted selection trait for body weight. These results provide a theoretical basis for selectivebreeding of P. olivaceus at different developmental stages.