Abstract: Twinning-induced plasticity (TWIP) steel, as an advanced high-strength steel, has great potential in the automotive field due to its high strength and excellent ductility. Cold rolling is a key method to improve its strength; the evolution of microstructure, changes in texture, and deformation heterogeneity during its deformation process have a significant impact on its properties. However, existing experimental methods have limitations in dynamic tracking and characterization, so it is necessary to use simulation methods for in-depth exploration. Therefore, in this paper, a phenomenological crystal plasticity model considering the twinning mechanism is used to simulate the cold rolling behavior of TWIP steel. By introducing mesh remeshing technology, the numerical instability problem caused by mesh distortion is effectively solved. The texture evolution law of the cold rolling process obtained by crystal plasticity simulation is consistent with the experimental results. The crystal plasticity simulation shows that during the cold rolling deformation of TWIP steel, regions with high kernel average misorientation (KAM) values are characterized by a relatively high proportion of Goss texture and relatively low proportions of Brass and Copper textures, while the opposite is true for regions with low KAM values. The study further found that the degree of grain deformation during the rolling process exhibits three typical modes: significant deformation heterogeneity inside the grains leads to their fragmentation and scattered orientations; minor deformation heterogeneity allows the grains to remain intact with concentrated internal orientations; and the grains split into several sub-regions with uniform deformation and concentrated orientations, where there are significant orientation differences between the sub-regions and grain refinement occurs. In addition, there is a regular rotation of grain orientations during the rolling process: most grains rotate such that their \bar\text1 , 1, 1 and 0, 0, 1 directions are parallel to the rolling direction (x). In the inverse pole figure, the final orientations are concentrated near the line connecting these two directions, among which the \bar\text1 , 1, 1 direction exhibits the highest orientation density; in the direction parallel to the rolling normal direction, the final orientations of grains in the inverse pole figure are concentrated near the line connecting the 0, 1, 1 and \bar\text1 , 1, 2 directions. In summary, by combining the crystal plasticity finite element method with mesh remeshing technology, this study systematically reveals the dynamic evolution law of deformation heterogeneity during the cold rolling process of TWIP steel, clarifies the evolution laws of microstructures (such as grain orientation and texture components), and provides a reference for understanding the correlation between the plastic deformation of materials and performance regulation.