The droplet growth processes during dropwise condensation are simulated with the help of computer. We focus on instantaneous and time-averaged characteristics of droplet size distributions. Based on the simulation results, the shift of a single peak from a small to a large size represents a significant characteristic for the instantaneous distribution before the first departure. Coexistence, shift, and combination of multiple peaks are the dominant features when condensing surface was refreshed repeatedly by shedding droplets from place to place. This indicates that the instantaneous droplet size distribution depends heavily on the growth time and target area. These findings help to explain why different distribution characteristics were reported in the experiments. Unlike the instantaneous distribution, time-averaged size distributions for coalesced droplets follow a power-law style due to a collaboration of coalescence events and renucleation behaviors. However, the size range for the power-law distributions was affected by nucleation density, which requires an appropriate usage of the empirical or fractal model to theoretically predict the heat transfer rate of dropwise condensation. This work provides a comprehensive understanding of the instantaneous and time-averaged characteristics of droplet size distributions.