ABSTRACT.- Residual feed intake (RFI) is a relevant trait, but expensive to measure so few candidates are phenotyped and have accurate GEBVs. Genetic evaluation corresponds to a single trait evaluation, where RFI are the residual from a lineal regression on feed intake on performance traits during feed-test. This study evaluates two approaches to predict a GEBV for RFI. First combing feed performance traits in a multitrait model to calculate a single measure of RFI using selection index, and 2nd the inclusion of weaning weight (WW) as an indicator trait combined with previous model. The GEBVs for the alternatives proposed were equivalent to conventional RFI. Using WW as a predictor trait on reference animals and candidates could be an inexpensive way to increase the accuracy of GEBVs. It allows to remove possible bias due to preselection of animals that participate on feed test and to extend prediction to more selection candidates.