The objectives of this study were to evaluate the relationship between live measurements and carcass traits and use the develop linear regression models to predict live weight and some of carcass traits in the local Black Baladi (BB), White Nicholas (WW) turkey strains and their repeated backcrosses according some body measurements at early age. Pearson's correlation was used to determine the coefficient of simple correlation between live weight, body measurements and the target carcass components (carcass weight and edible parts). Stepwise multiple regressions were performed to estimate live weight and carcass weight at 20 wks of age using both of body weight and body measurements traits at 16 wks of age to produce the best regression model for each of the dependent variable based on the regression coefficient. Results obtained from descriptive statistics showed that the differences of mean values among the different genotypes, live weight at 16 and 20 week of ages (BW16 and BW20), shank length (SL), keel length (KL), breast width (BW), breast circumference (BC) were highly significant (P<0.01) and influenced by repeated backcrosses. This was also applicable to carcass (CW), and edible parts (EP) weights. Simple Pearson correlation coefficients (r) between body weight at 20 wks of age and body measurements (SL, KL, BW and BC) and carcass yields (CW and EP) had positive and significantly high values of most of the studied traits where (r = 0.25 to 0.99) for the four genotypes (except BW for two repeated backcrosses which had negative and low values). The results of stepwise multiple regression reveals that BW16 seems to be the major trait in determining for predicting BW20, CW and EP base on high adjusted determination coefficient (R2) as shown in all equations. These results based on R2 change for each independent variable. Generally, all models for predicting the former three dependent variables had highly significant .High coefficient of multiple regression between the dependent and independent variables and consequently, high R2 and adjusted R2 values (P<0.01).