<p>Efficient development of heterogeneous reservoirs remains challenging due to early channeling and poor sweep efficiency during conventional and chemical flooding. To address this issue, a pore-scale numerical framework coupling the Navier–Stokes and Discrete Element Method was established to systematically evaluate the effects of particle elasticity, particle size, and slug configuration on enhanced oil recovery. The model captures the coupled dynamics of multiphase flow, particle deformation, and interfacial transport, enabling quantitative analysis of the interaction mechanisms within heterogeneous composite flooding systems. Results show that particle elasticity and size jointly control flow diversion and displacement efficiency. High-elasticity particles enhance flow redistribution and pressure buildup, whereas low-elasticity particles improve injectivity. Large particles mainly form macroscopic plugging zones, while small particles enhance pore-scale sweep, and their combined injection achieves effective coupling between profile control and microscopic displacement. In composite heterogeneous–chemical systems, recovery improvement results from the synergy between elastic plugging, small-particle deep displacement, and polymer–surfactant interfacial regulation. Among all tested configurations, the high-elasticity large-particle / low-elasticity small-particle / high-elasticity large-particle sequence produced the lowest residual oil saturation. This study provides pore-scale insights for optimizing slug design in heterogeneous reservoirs.</p>