The selection of ideotypes combining high biomass production, high water use efficiency (WUE) and adequate wood properties is a major challenge in improving forest trees for industrial plantations. This issue was addressed in a field experiment evaluating 1130 clones of Eucalyptus urophylla * Eucalyptus grandis. Genomic selection using 3303 SNPs and the GBLUP model was carried out at age 55 months to select genitors and clones for stem volume (V55), WUE (using stable carbon isotope composition in stemwood, d13C as a proxy, transformed into intrinsic WUE, Wi), as well as lignin (KL) and holocellulose (HCEL) contents. The variance components were mainly additive for d13C(Wi), KL and HCEL, while they were highly non-additive for V55 (the ratio of dominance to additive variance was 130%). The narrow sense heritability was higher for Wi (h2 = 0.704) than for volume and wood traits (<0.5), showing that this trait was under strong genetic control. Additive and total genetic correlations among traits were low (between -0.260 and 0.260) as were the correlations due to environmental effects (between -0.267 and 0.344), which suggests a rather independence among traits. The equal emphasis and the desired gain index selection methods were used to assess economic weights and to estimate the expected additive and total genetic gains. Different breeding objectives were defined with percentages weighting the economic coefficients or the desired gain for the equal emphasis or desired gain method, respectively. When the breeding objective gave emphasis to volume production using the percentages of 70%, 10%, 10%, and 10% for V55, HCEL, KL and Wi, respectively, positive gains were observed in both methods, but the correlation between index rankings were 0.686 and 0.635 for breeding and clonal selection, respectively. With a more balanced breeding objective using percentages such as 25%, 25%, 25%, 25% for V55, HCEL, KL and Wi, respectively, both index methods gave close positive gains an