This paper attempts to model the main physicomechanical properties of a wood sand concrete using the 'design of experiments' method. The limestone filler content (F), the maximum diameter of sand grain (D), and the wood content (B) were taken as independent variables; while the dry density (d), the thermal conductivity (C), and the compressive strength (Rc) were taken as responses. For this study, experimental tests were performed and the above method was applied. The obtained results showed that the most significant variable that affects all the studied responses is 'B.' The effect of the other factors is relatively weaker with very clear domination of the factor 'D' compared to the factor 'F' which seems to be very negligible. Although it is also lower than those of 'B' and 'D,' the effect of the binary and ternary interactions (DxB, FxB, FxD and FxDxB) seems to be slightly higher than that of (F). In addition, the SEM analysis highlighted the more or less homogeneous aspect of the studied composite as well as the good adherence 'wood-matrix' and confirmed the dominant effect of the factor 'B' on the microstructure evolution of sand concrete. Moreover, according to 'B' it is possible to develop both structural concretes and structural-insulating concretes.