As a common natural phenomenon, icing is widespread in industrial production processes and often causes adverse effects on daily life and production. The potential hazards caused by icing in various industries could be significantly reduced or even eliminated by using ice characteristics prediction technologies. Previous studies have shown that the rate of road traffic accidents can be reduced by 65% after using ice characteristic prediction technology. Furthermore, when combined with the de-icing system, 80% of icing scenarios can be anti-icing or de-icing. To effectively address the issue of ice disasters across various industries, this review provides an overview and analysis of existing icing prediction technologies based on stationary surfaces such as road and transmission line, as well as moving surfaces such as wind turbine blade and aircraft. The results indicate that existing technologies achieve prediciton accuracies of over 80% for indicators such as ice thickness on stationary surfaces and over 70% for moving surfaces. Existing ice characteristic prediction technologies can be divided into two types, model-driven method and data-driven method, with the later showing significant potential for development. Based on the summarized ice prediction technologies applied to four simple cold surfaces under stationary and moving states, this paper further proposes key research directions in this field, aiming to provide reference and guidance for the development and optimization of anti-icing technologies on low-temperature surfaces in various engineering scenarios.