The selective separation of ions in aqueous solutions is of great importance for water purification and resource recovery. Capacitive deionization (CDI) has emerged as a novel electro-driven desalination process and is gradually being applied in the selective separation of target ions from multicomponent solutions. By leveraging customizable electrode materials, interfaces, and adaptable operating conditions, CDI offers substantial advantages in modulating ion adsorption processes, thereby enabling the selective removal of specific species. In this review, we examine the developmental trajectory of CDI, with a focus on the selective mechanisms of innovative materials and operational influences. For carbon-based capacitive materials, unmodified electrodes achieve selectivity through ion properties such as hydration radius matching (e.g., microporous activated carbon cloth prioritizes Cs+ over Na+/Ca2+), valency (higher-valency ions are preferred via stronger electrostatic interactions), diffusion rate (ClO4− exhibits ~11-fold selectivity over Cl− due to faster pore diffusion), and electronegativity (ReO4− > NO3− > Cl− via hydrogen bonding under pH=2). Chemical modifications enhance selectivity: CTAB/SDBS-functionalized carbon electrodes exhibit 7.7-fold higher adsorption of NO3− over Cl−; amino-modified activated carbon increases SO42− selectivity by 1.98-2.52 times; EDTA-graphene selectively captures Pb2+ via chelation, allowing stepwise desorption to separate Pb2+ from Na⁺. For pseudocapacitive materials, Ni-Al-LMO achieves ~50 mg g−1 F− adsorption (4 × that of Cl⁻); Na0.44MnO2 shows Na+/K+ selectivity of 13 and Na+/Ca2+ selectivity of 6-8; CuHCF removes 80% of NH4+ with > 4× selectivity over Na+, while NiHCF prioritizes NH4+ > K+ > Na+ via intercalation. Redox-active PVF-CNTs achieve selectivity values >40 in aqueous and >3000 in organic systems for organic anions. Regarding ion-selective membranes, CIMS CEM shows Na+/Ca2+ selectivity of 1.8; polyelectrolyte-modified AEMs enhance divalent anion adsorption; PSS-coated electrodes reach a Ca2+/Na+ selectivity coefficient ~8; CSN-coated electrodes increase Ca2+ adsorption by 50.7% compared to uncoated ones; nitrate-selective resins raise NO3−/Cl− selectivity to 3.7. Operational parameters also affect selectivity: lower voltages improve selectivity (e.g., NO3−/Cl− selectivity decreases with higher voltage) but at the expense of ASAR; higher target ion concentration ratios (e.g., elevated NO3−/Cl−) enhance selectivity; pH modulates H2PO4−/HPO42− adsorption and Cr(VI) reduction to Cr(III); and prolonged operation time induces ion displacement (e.g., Na+ initially adsorbed is later replaced by Ca2+). Finally, this paper highlights several challenges: limited selectivity of traditional carbon electrodes, the need for novel membranes, incomplete performance evaluation (e.g., lack of standardized adsorption capacity metrics based on active components, neglect of energy efficiency and product purity), insufficient system design, poor material stability (e.g., lack of long-term data, toxicity of metal oxides), and gap between simulated and real wastewater applications. Future prospects include expanding CDI applications to the separation of organic/neutral substances (e.g., proteins via pH tuning), applying machine learning for material selection and parameter optimization, and advancing techno-economic assessments in the electronics, rare earth, and food industries, thereby providing guidance for the development of low-carbon, high-efficiency CDI processes.