The recovery of valuable materials from spent lithium-ion batteries (LIBs) has experienced increasing demand in recent years. Current recycling technologies are typically energy-intensive and are often plagued by high operation costs, low processing efficiency, and environmental pollution concerns. In this study, an efficient and environmentally friendly dielectrophoresis (DEP)-based approach is proposed to separate the main components of "black mass" mixtures from LIBs, specifically lithium iron phosphate (LFP) and graphite, based on their polarizability differences. A custom-designed microparticle separator is developed for the continuous separation of LFP and graphite mixtures at high throughput. Additionally, a theoretical model incorporating both electric and flow fields is constructed to predict the DEP behavior of particle streams. The feasibility of selective separation is theoretically evaluated through numerical simulation of microparticle trajectories and binary separation within the proposed separator, and these results are experimentally validated with good agreement. Under a particle streamflow rate of 10.8 mL/min, both simulations and experiments demonstrate a separation efficiency of LFP higher than 80% at 100 V. Furthermore, the influence of operating parameters, such as the applied voltage, flow rate, and sheath-to-feed ratio, on optimal separation efficiency and particle purity is numerically investigated. The feasibility of the proposed separator for the potential separation of other lithium-metal-oxide-containing particle mixtures is also explored through numerical simulations. Overall, this study provides a theoretical foundation for the development of high-performance and sustainable LIB recovery processes with a low energy consumption.