This study presents a numerical model for analyzing the impact of key factors influencing black ice formation and proposes a prediction model based on the numerical simulation results. Black ice, once formed, can significantly reduce pavement friction, thereby increasing the risk of traffic accidents. Although experimental measurements can be used to investigate the effects of various parameters on black ice formation and to develop predictive models, their application in real-world scenarios is limited. This limitation is primarily due to restricted access to high-traffic roads for continuous data collection, as well as the time-consuming and costly nature of conducting a large number of experiments. To overcome these challenges, a detailed parametric numerical analysis was conducted to assess the effects of ambient air temperature, wind speed, pavement temperature, and water film thickness on black ice formation. Based on the results of these simulations, a dataset comprising 1800 different scenarios was generated and used as input for a logistic regression model developed to predict black ice formation. The performance of the proposed model was validated using an experimental dataset consisting of meteorological data, pavement surface temperature, and ice thickness measurements collected from an expressway segment in South Korea. The validation results demonstrated satisfactory accuracy in predicting black ice formation.