Soil-woody biomass interactions are relevant for the productivity of bioenergy plantations and biomass quality. In this context, the main objective of this study was to evaluate and to quantify possible relationships between chemical variables of the soil and the produced biomass through a multivariate approach. This latter approach allows to overcome the complex issue of multi-collinearity among variables. Soil and woody biomass samples were collected from two poplar short rotation coppices in Santarem (Portugal) and in Lochristi (Belgium). The results from the analyses of those samples were integrated into three databases with soil, woody biomass and site plots as cases, and 23 physical and chemical properties as variables. The databases were subjected to a multivariate sequence of calculations, which included correlation, principal components, factorial and hierarchical clustering analyses. The calculations showed that the site plots and the woody biomass of genotypes in Lochristi were more homogeneous as compared to Santarem; they also confirmed the high interconnection between soil and woody biomass variables. The higher heating value of the woody biomass correlated well with the soil concentrations of P2O5, Mg, Ca, Na and organic C. Linear equations related the higher heating value to the most important soil and woody biomass variables. Finally, the results suggest that the annual monitoring of soil and biomass in SRC systems should be performed to optimize both productivity and woody biomass quality as a fuel. (C) 2017 The Authors. Published by Elsevier Ltd.