Abstract: Widespread micro-terrain microclimates pose a serious risk of icing for transmission lines. However, the current definition of micro-terrain remains a vague descriptive definition, and existing quantification methods mainly rely on a large number of observational samples and empirical knowledge. This paper proposes a method for extracting indicators and identification criteria for micro-terrain based on digital elevation models (DEMs) and statistical data on icing disasters. Simplified geographical models are constructed based on geographical feature boundaries for various micro-terrain types. Validation of the classification and identification conditions is performed through numerical simulation calculations and wind tunnel experiments. Moreover, the meteorological characteristics are simulated under different micro-terrain conditions, and the reliability of the identification conditions for ice-covered micro-terrain is verified. The results indicate that icing micro-terrain is typically distributed continuously and is less likely to exist in isolation, often exhibiting complex combinations of multiple micro-terrain types. Typical icing micro-terrain that meets the identification criteria displays distinct micro-meteorological characteristics, with the most pronounced effects at windward entrances and maximum wind acceleration occurring at mountain summits except for saddle types. This method overcomes the shortcomings of clustering methods that lack geographical interpretability and reduces sample dependence, providing design standards for transmission lines in the absence of icing observations.