Carcass data are nowadays easy collected in routinely regime from Dutch slaughter houses. The aim of this study was to develop a selection index for beef production traits in a dairy cattle population based upon such data. Records during 10 years (1995-2004) for body weights at 12 (w12), 18 (w18) months of age and four years (2001-2004) for hot carcass weight (HCW) and lean weight (LW) at the experimental farm of Faculty of Agriculture, Minufiya University were utilized to construct different selection indices (general index, reduced indices, and sub-indices) by using multisource of information (W12 and W18 as an Own-performance traits ; LW and HCW as a Paternal half-sibs traits) to improve some beef characteristics in Frisian bull calves. The secondary objective is to evaluate and predict genetic parameter estimates of body weights at 12, 18 months of age, HCW and LW. Overall means for the previous body weights were 291.97, 358.73, 283.54 and 215.09 kg respectively. Heritability estimates for the previous traits were 0.59, 0.71, 0.67 and 0.29, respectively. All estimates of genetic (rG) and phenotypic (rP) correlations among different body weights were positive. Fifteen selection indices were constructed using four traits and two sources of information in different combinations, own-performance for (w12), (w18) and Paternal half-sibs for (HCW) and (LW), indices (I2), (I3) , (I5), (I6), (I9), (I10) and (I13) gave high (RIH) and (RE) values compare with general index (I1). Therefore, it could be suggested that to use (I3) and (I1) to improve beef traits in Friesian bull calves under the large scale because the highest values of excepted genetic change for lean weight as an econo