Mesenchymal stem cell (MSC)-based regenerative therapies are promising for healing tendon injuries and tears, due to their potential to differentiate into tenogenic cells. However, generating homogeneous populations of tenogenically differentiated stem cells remains a big challenge, as non-differentiated cells can lead to post-transplantation complications. Therefore, a homogenous sample of tenogenically differentiated MSCs is critical for advancing tendon therapies and avoiding uncontrolled cell growth or non-tendon tissue formation (e.g., ectopic bone). This work is focused on designing and simulating a dielectrophoretic (DEP)-based label-free, microfluidic platform to selectively sort and enrich tenogenically differentiated MSCs (tMSCs) from undifferentiated MSCs. Using particle tracing, creeping flow (transport of diluted species model), and electric current physics modules in the COMSOL Multiphysics simulation software package, the sorting was simulated within a two-stage microfluidic device operating at a sinusoidal frequency of 160 kHz. The optimal separation efficiency and purity are achieved at an inlet velocity of 400–1000 μm/s, with specific voltage configurations, enabling recovery of one million tMSCs in ∼3 h. Results demonstrate a near-linear relation between recovery time and particle count at the outlet boundaries and selected surfaces, indicating consistent throughput across varying conditions. This study demonstrates that DEP can offer a scalable, efficient, and label-free method for enriching tMSC populations with high selectivity, enhancing more prospects for MSC-based tendon therapies and advancing the development of microfluidic sorting devices for regenerative medicine applications.