Ice accretion is a phenomenon that forms an ice layer on a solid surface when supercooled droplets or ice particles in the atmosphere impinge there. Ice accretion on an aircraft threatens navigation safety because it changes the wing shapes and reduces aerodynamic performance. Hence, it is a significant technical issue to predict icing during the design phase of an aircraft. In the glaze ice conditions, where the ambient temperature is about −10 to −3 ℃, the impinged supercooled droplets do not freeze instantly but runback along the wing surface as a liquid film flow. In addition, splashing and rebounding occur, which generate secondary droplets if supercooled large droplets (SLD) whose diameters are over 40 µm impinge on an aircraft. Due to these secondary droplets, very complex ice shapes are formed in SLD icing conditions. Therefore, the prediction of SLD icing under glaze ice conditions is very difficult in terms of the complexity of physics. In the present study, the icing simulation by the hybrid scheme composed of grid- and particle-based methods was conducted to improve the accuracy of ice shape prediction. First, the flow field and the droplet trajectory were computed using the grid-based method. When the droplets approached the airfoil or ice surfaces, behavior of droplets and heat transfer were simulated by the particle-based method. In this study, we used the Explicit-Moving Particle Simulation (E-MPS) method of a particle-based method with the governing equations of incompressible Navier-Stokes, energy, and continuity equations. This sequential process was repeated until a specified termination time in multi-shot simulations. After every 30 seconds of the icing simulation, the flow field was updated with Icing Cell Method (ICM) to account for changes in the droplet trajectory due to the ice accretion. As a result, the ice mass on the pressure side decreased in the multi-shot simulation compared to the single-shot simulation. Although feather-like ice shapes were observed in both methods, the ice shape of the multi-shot simulation is closer to the experimental data than that of the single-shot simulation. Through this study, it is indicated that the hybrid method is useful in simulating icing under glaze ice conditions.