AbstractHybridization in the Urochloa humidicola breeding programme allowed to explore the genetic variability of the apomictic sources and, thus, to obtain new sexual and apomictic parents with compatible ploidy and/or superior progenies. However, given the high variability arising from crosses, there is a need to adopt efficient selection strategies among and within progenies. The aim of this study was to compare the selection methods among U. humidicola progenies with (t‐BLUP) and without (BLUP) the weighting based on prediction error and subsequent selection within progenies by individual simulated (BLUPIS) or optimal individual BLUP. We evaluated seventy‐one U. humidicola full‐sib progenies from biparental crosses between nine sexual and ten apomictic parents. The experiment was set up in an incomplete block design with 64 blocks of 15 plots each, consisting of individual plants. Seven consecutive cuts were made and the traits total dry‐matter, leaf dry‐matter and regrowth scores were measured. Genetic variance was expressive for all traits. The selection procedure among progenies of U. humidicula based on t‐BLUP proved to be more advantageous regarding the use of BLUP not weighted by the prediction error, resulting in higher selection gains, taking as reference the optimal procedure of the individual BLUP. Furthermore, the use of BLUPIS allowed an optimized selection of hybrids, allowing the breeder to explore with variable selection intensity the genetic variability within the selected full‐sib progenies based on relative genetic merit.