High-throughput analytical techniques to characterize cell wall composition in large-scale lignocellulosic biomass samples from a diverse set of clones / genotypes are essential for unravelling the underlying genetic architecture of wood properties, such as lignin, cellulose and hemicellulose. This, in turn, accelerates the genetic improvement of cell wall composition to reduce biomass recalcitrance during biofuel production. For these purposes, there is an increasing interest in combining Near-infrared (NIR) spectroscopy and multivariate statistical analysis. In this study, we investigated the potential of using this method to predict wood properties, with the aim of applying the predictions to understand the genetic basis of these complex traits in black poplar (Populus nigra L.). Populations representing the natural range of the species in Western Europe were grown in two clonal trials at two contrasting sites under a short rotation coppice (SRC) system and NIR spectra were collected from ca 6,000 wood samples. A subset of 120 calibration samples covering the entire range of spectral variation in the whole population were selected and analyzed using standard methods (wet chemistry, high-performance liquid chromatography, analytical pyrolysis) for wood properties. The absorption spectra and reference values of calibration samples were then employed to develop calibration models at a global scale using partial least squares (PLS) regression and crossvalidation. Global models for predicting extractives, C5 / C6 ratio, S / G ratio, xylose / glucose ratio, Pylignin and soluble lignin had pretty high coefficients of determination (R2cv : 0.71 – 0.86), while the model quality for Klason lignin and common wood sugars (glucose, xylose) was moderate (R2cv : 0.50 – 0.68) and strongly varied depending on the site or coppice rotation considered. The correlations between wood chemical traits and the potential application of NIR analysis for predicting glucose yield following saccharification have also been studied. Our results show the promise of using NIR spectroscopy calibration models for high-throughput phenotyping of wood properties in natural populations of black poplar. Genetic analysis with wet chemistry and analytical pyrolysis data in the 120 calibration samples revealed that the studied wood properties were under moderate genetic control (broad sense heritability ranging from 0.24 to 0.66). The extension of such analyses to the NIRS predicted values in the entire population and their application to association mapping are also discussed.