The proliferation of electric vehicles and stationary grid storage, as well as the ongoing transition to renewable energy generation technologies, require continuing improvements to current lithium-ion batteries (LIB) in terms of energy density and lifetime performance. Silicon (Si) has been considered as one of the most promising replacements for graphite in LIB anodes due to its superior energy density (up to 10 times higher than that of graphite). However, one prominent issue with Si-based anode materials is their large volume change upon lithiation and delithiation, leading to particle pulverization, electrical isolation, and unstable formation of the solid-electrolyte interphase (SEI) that consumes lithium (Li) ions irreversibly. In this thesis, we report several approaches to advancing Si-based anode design for next-generation high-capacity LIBs. First, a low-cost method is introduced for producing a composite consisting of Si particles and pyrolyzed polyacrylonitrile (PPAN) as a conductive binder, resulting in an anode with high cycling stability of 1000 mAh/g for over 800 cycles in half-cells. We then combine experiments with ReaxFF simulations to investigate the structure–property characteristics of PPAN, as compared to other polymeric binders, in dictating the performance of Si/binder anodes. PPAN is determined to be a highly promising binder for Si anodes due to its structural orientation at and near the interface, enabling good adhesive strength, mechanical elasticity, and Li-ion conductivity. This work also showcases the effectiveness of combining modeling with experiments as a synergistic approach for binder evaluation, selection, and optimization. Second, we present the successful replacement of copper current collectors with non-woven carbon nanotube fabrics (CNTf) to combine with our high-capacity Si/PPAN material, resulting in a self-standing anode construction that is truly lightweight, potentially boosting the cell-level energy density by 25% when paired with commercial cathodes. Excellent interfacial contact between the Si/PPAN layer and the underlying CNTf is also revealed, as well as the effects of varying CNTf thickness. Overall, our results have the potential to expand the use of CNTs in high-energy LIBs and contribute to the reduction in greenhouse gas emissions associated with metallic component manufacturing. Third, we demonstrate the application of stabilized lithium metal particles (SLMP) as a pre-lithiation method for mitigating the irreversible Li loss associated with Si anodes in full-cell LIBs. Particularly, a surfactant-stabilized SLMP dispersion is designed to be spray-coated onto our Si/PPAN anodes, forming a homogeneously distributed and well adhered SLMP layer for in-situ pre-lithiation. The effects of SLMP at high and low loading ratios on the Si/PPAN in full-cells under different cycling regimes is then closely examined. Our findings provide valuable insights into the design of pre-lithiation and cycling strategies for high-capacity Si-based full-cell LIBs, in order to fully utilize the benefits of SLMP while avoiding the undesirable Li trapping effect.