With the growing demand for the exploration of unconventional oil and gas resources, the drilling industry is progressively advancing into the era of deep and ultra-deep wells. During the drilling of 10,000-m-deep well, the diameter of the upper casing is gradually increased to ensure adequate operational space for the lower sections. As the casing diameter increases, the annulus fluid velocity decreases, leading to reduce of the cuttings-carrying capacity. Furthermore, the cuttings flow patterns are complex and highly variable, their transition mechanisms remaining unclear. In this study, the CFD-DEM method was used to simulate the cuttings transport in the upper large-scale annulus during ultra-deep well drilling. Firstly, the model validation was confirmed through grid independence test. By referencing actual field drilling parameters, this study investigated changes in wellbore pressure drop, cuttings flow patterns, and annulus flow behaviors under varying parameters, e.g., fluid velocity, drillpipe rotation speed, and dimensionless drillpipe diameter. The results showed that fluid velocity is the primary factor influencing hole-cleaning efficiency. Additionally, it found that as the drillpipe rotation speed, dimensionless drillpipe diameter and fluid velocity increase, the flow patterns transition into clustered, clustered-spiral, spiral and homogeneous flow. Through microscopic characterization of cuttings mechanical properties, the mechanism behind these flow patterns transitions was revealed. This study referenced total Reynolds number, combined with changing of dimensionless pressure drop, a new cuttings flow pattern plate has been developed. The findings of this research offer valuable theoretical insights for optimizing field drilling parameters.