coal powder sedimentation and blockage represent one of the principal constraints on the efficient production of coalbed methane (CBM) well. The initial liquid-only flow stage during dewatering represents the main period of extensive coal powder production, and analyzing the settling behavior of coal powder is essential for optimizing its performance. Most existing settling models assume spherical particles and employ two-dimensional shape parameters, which fail to represent the three-dimensional heterogeneity of particles, resulting in a limited predictive accuracy across flow regimes. In this study, single-particle sedimentation experiments, scanning electron microscopy (SEM), and three-dimensional laser profilometry were conducted. The Wadell sphericity and axial ratio parameters were integrated to construct a composite shape factor (ψ) and to fit an equivalent diameter. The ψ factor was then incorporated into the Ferguson–Church model, in which the drag coefficient parameters (C₁, C₂) and the flow regime index (n) were optimized through segmented calibration and least-squares fitting, thereby enabling accurate characterization of the settling behavior of non-spherical coal powder particles. The experimental results and residual analysis indicated that the improved model achieved markedly enhanced predictive accuracy across all flow regimes(R²˃0.95, mean error ≈ 4.2%). Furthermore, COMSOL numerical simulations were used to elucidate the mechanisms through which particle size and geometry influence the flow field structure and settling trajectories. These findings provide a quantitative foundation for modeling coal powder transport, designing wellbore anti-blocking strategies, and optimizing drainage operations, thereby offering important theoretical and engineering implications.