We investigate various lattice defects using electronic-structure-based first-principles calculations to evaluate the factors governing material strength that are strongly influenced by elemental properties. Furthermore, we develop machine-learning models trained on first-principles calculation data to reveal lattice defect dynamics and related properties in large-scale atomistic models. To evaluate nanoscale deformation mechanisms and the effects of stress and temperature, we employ a wide range of computational approaches, including energy-based methods and dynamical analyses. By integrating these approaches with data-driven methods and nanomechanical testing, we conduct comprehensive evaluations and predictions of the factors governing the strength of metallic materials. We also gain new insights into the formation processes of microstructures in crystalline and amorphous materials that influence their mechanical strength through atomistic analyses. By directly evaluating nanoscale microstructure formation and strength-controlling factors, we aim to establish fundamental, physics-based principles for improving the mechanical properties of a wide range of materials.