In this study, the complex two-phase gas–solid flow inside an industrial-scale classifier is successfully captured using two-way coupled CFD-DPM in Ansys Fluent R2 2022 with the assistance of a High-Performance Computational System. The flow field inside the classifier, characterized by multiple vortices with different positions, sizes, shapes, and rotational directions, is significantly affected by rotor speeds. This directly impacts material particle flow and classification performance, evaluated by two important conditions: classification efficiency (η) and particle size distribution curve, along with the common classification indexes (d50, K). As the rotor speed increases from 240 to 300 rpm, Newton’s efficiency gradually increases from 88.3% to 95.74%, while d50 and K decrease by about 22% and 11%, respectively. The particle size distribution of the final product was significantly changed under different rotor speeds. Additionally, simulation analysis of the structure and velocity distributions of the flow field inside the full-scale industrial classifier helps comprehensively understand the effects of rotor speeds on classification performance and the classification mechanism. This provides a foundation for the optimal design and selection of optimal parameters for the classifier to achieve the required classification performance.