Traditionally, comprehensive laboratorial experiments on newly proposed microfluidic devices are necessary for theoretical validation, technological design, methodological calibration and optimization. Multiple parameters and characteristics, such as the flow rate, particle size, microchannel dimensions, etc., should be studied by controlled trials, which could inevitably result in extensive experiments and a heavy burden on researchers. In this work, a novel numerical model was introduced to simulate particle migration within a complicated double-layered microchannel. Using the hybrid meshing method, the proposed model achieved a significant improvement in meshing quality, and remarkably reduced the required calculation resources at the same time. The robust, efficient and resource-saving numerical model was calibrated and validated with experimental results. Based on this model, 1) the mechanism of microparticle manipulation within the microchannel was revealed; 2) the primary reason for the microparticle focusing failure was investigated; and 3) the optimal microparticle sorting strategy at different flow rates was analyzed. In experiments, the obtained optimal strategy could approach a good sorting performance with a high recovery rate and high concentration ratio in a high-throughput manner. The proposed numerical model shows great potential in mechanism investigation and functional prediction for microfluidic technologies using unconventional designs.