Abstract: Following the random characteristics of ice accretion process, a prediction method is proposed to investigate the density of ice accretion under different atmospheric conditions. Based on the property of droplet motion, a random walk model for the droplets along ice structure has been introduced to simulate the ice accretion process. As a result, the effects of model parameters including freezing and motion parameters on ice density could be obtained. In addition, a relationship between atmospheric factor and two parameters of random-walk model has been established. Consequently, different types of ice structure and density could be predicted. Simulation results show that the ice density and accretion structure predicted by introduced random walk model agree qualitatively with experimental observations.