Wood density affects the strength of lumber and paper products. Despite considerable research, however, the key factors influencing wood density are still not fully understood. The aim of this study was to analyze the effects of variables within and among trees on the density of current wood growth along the stem in order to further understanding and prediction of wood quality. Correlation analyses together with linear mixed effects and randomForest modelling were carried out using stem analysis and tree-level data from long-term balsam fir and white spruce stand density experiments in Eastern Canada. Wood density showed highly structured patterns of variation among trees. Mixed effects and randomForest models, which incorporated tree-level effects accounted for much of the wood density variation. The three most influential variables identified by randomForest analysis for both species were tree slenderness, relative height and ring width. Wood density increased with slenderness and relative height, and these variables affected relationships of wood density to ring width. Wood density is associated with interacting variables within and among trees including tree slenderness and relative height. Wood density is related to ring width, but this relationship is modulated by tree-level influences which likely reflect mechanical stability requirements.