Two Types of Ice Clouds (TICs) have been characterized in the Arctic during the polar night and early spring. TIC-1 are composed by non-precipitating small ice crystals of less than 30 µm in diameter. The second type, TIC-2, are characterized by a low concentration of large precipitating ice crystals (>30 µm). Here, we evaluate the Weather Research and Forecasting (WRF) model performance both in space and time after implementing a parameterization based on a stochastic approach dedicated to the simulation of ice clouds in the Arctic. Well documented reference cases provided us in situ data from the spring of 2008 Indirect and Semi-Direct Aerosol Campaign (ISDAC) campaign over Alaska. Simulations of the microphysical properties of the TIC-2 clouds on 15 and 25 April 2008 (polluted or acidic cases) and TIC-1 clouds on non-polluted cases are compared to DARDAR (raDAR/liDAR) satellite products. Our results show that the stochastic approach based on the classical nucleation theory, with the appropriate contact angle, is better than the original scheme in WRF model to represent TIC-1 and TIC-2 properties (ice crystal concentration and size) in response to the IN acidification.