The conventional fill compaction procedure usually controls the dry density ρ_d and the water content w referring to the maximum dry density (ρ_d)_max and the optimum water content w_opt determined by laboratory compaction tests performed at a certain compaction energy level (CEL). However, (ρ_d)_max increases and w_opt decreases with CEL, while, since Proctor (1933), CEL practically available in the field has been increasing and the required ρ_d value has generally been becoming higher for more satisfactory performance of soil structures. Besides, the values of (ρ_d)_max and w_opt change with soil type. In a single earthwork project, the actual CEL and soil type may vary and it is very difficult to accurately estimate the field CEL and identify the actual soil type at a given moment at a given place. Therefore, the actual values of (ρ_d)_max and w_opt are usually unknown. On the other hand, the optimum degree of saturation (S_r)_opt defined as S_r when (ρ_d)_max is obtained for a given CEL and the ρ_d/(ρ_d)_max vs. S_r−(S_r)_opt relation are rather independent of CEL and soil type. The unsoaked and soaked strength and stiffness and the saturated hydraulic conductivity are a function of ρ_d and compacted S_r not including CEL as a variable. It is proposed to control S_r of compacted soil to be close to (S_r)_opt and ρ_d to be large enough to achieve the physical properties required in design, together with pre-compaction control of water content.