Objective: Icing on overhead transmission lines presents a serious risk to the safety and stability of power grids, particularly amid the rapid expansion of ultra-high voltage networks in China. Accurate simulation of conductor icing is crucial for effective disaster prevention and mitigation. However, many existing models primarily depend on the median volume diameter (MVD) of droplets, often overlooking droplet size distribution (DSD) characteristics and leading to simulation errors. This study addresses this challenge by developing an optimized method for selecting droplet size parameters in icing simulations, thereby improving computational accuracy and efficiency. This work is essential for enhancing the reliability of icing predictions and reinforcing the resilience of power infrastructure under extreme weather conditions. Methods: This study employs two primary methodologies: the analytical and Eulerian methods. The analytical method, based on the Finstad's model, calculates the droplet collision coefficient (α₁) using MVD, whereas the Eulerian method leverages computational fluid dynamics to simulate air-droplet two-phase flow, incorporating DSD for higher precision. A comparative analysis of these methods is conducted to evaluate their efficiency and accuracy. In addition, this study investigates the impact of environmental parameters (wind speed, MVD, and conductor diameter) and droplet dispersion on α₁ errors. A dynamic selection strategy is proposed to determine when MVD could suffice or when DSD is necessary based on predefined error thresholds. Results: The key findings included the following: the analytical method outperformed the Eulerian method in computational speed but tended to overestimate α₁ due to unaccounted turbulence effects. Meanwhile, owing to the influence of DSD, directly using MVD to calculate α₁ in conductor icing simulations also introduced a certain error. The error diminished with higher wind speeds and larger MVD values. Using MVD alone introduced errors (Δα₁) in α₁ calculations, which exhibited a nonlinear trend: Δα₁ initially decreased to zero and then increased as MVD, wind speed, or conductor diameter increased. To avoid calculation errors in the conductor's α₁, one might consider using the DSD instead of MVD for computing α₁. However, this method involved significantly greater computational requirements and was therefore unsuitable for rapid assessment of conductor icing accumulation. This study identified critical thresholds where MVD can replace DSD without significant accuracy loss, optimizing computational resources. In detail, leveraging the high computational efficiency of the analytical method, Δα₁ was calculated. When this error was less than or equal to the maximum allowable error, MVD could be used in place of DSD, thereby achieving an optimal balance between computational efficiency and accuracy. The results were validated through case studies using experimental data from Pavlo et al. Conclusions: This study highlights the limitations of MVD-based icing simulations and underscores the importance of droplet dispersion characteristics. By integrating analytical and Eulerian approaches, this study provides a practical framework for dynamically selecting droplet size parameters, ensuring accurate and efficient icing predictions. The results show that, although MVD suffices under specific conditions, DSD is indispensable for scenarios involving highly dispersed droplets or smaller conductors. This study advances the field by offering a scalable solution for power grid resilience against icing hazards, with implications for academic research and industrial applications.