Genome-wide evaluation (GWE) is expected to bring substantial benefits to the genetic improvement of perennials like poplar. Firstly, because current phenotypic evaluation is lengthy, costly and complex for the species. Secondly, because identifying poplar clonal varieties rely on a costly built-up of large segregation families from which to draw the best candidates combining favourable alleles with additive and non-additive effects.The present study provide preliminary assessments of the feasibility of GWE in a highly monitored subset of the French black poplar breeding pedigree. The training subset comprised a factorial mating design involving 6 parents and 1 known grand-parents and 261 individual records. Recorded traits were growth in the form of tree circumference and height, and architecture assessments taken as branch angle. Patterns of environmental heterogeneity in the trials were carefully taken into account at the individual tree level, by implementing autoregressive and splinebased two-dimensional smoothing surfaces into the evaluation model. All individuals have been genotyped thanks to a recently developed 12K SNP chip, which produced in our pedigree 7812 usable SNPs. Missing genotyping values, which amounted to 1.7% of data, were imputed by the FImpute software. Several genomic best linear unbiased predictors (G-BLUP) models were applied using GS3 software.Results provided prediction abilities for each trait obtained from a cross-validation approach. Accuracies of these predictions differ according to architecture and heritability of these traits. The benefits of imputation and environmental heterogeneity adjustments are discussed at the light of these results, together with the impacts that a GWE based selection would have on the relatedness among candidates compared to a pedigree-based approach.