In many engineering projects, an important aspect is the depth of ground freezing, which is influenced by factors such as the type of material, thermal properties of the soil, water content, and climatic conditions including temperature, wind speed, precipitation, and solar radiation. The depth of frost penetration can be estimated using numerical or analytical methods, provided a large number of input data points are available. However, such data is typically not readily accessible. In such cases, similar accuracy can be achieved using simpler analytical and semi-empirical models calibrated from observations. Ground freezing, being a random process largely dependent on climatic conditions, which are also random, should be analyzed using probabilistic methods. The paper deals with the probabilistic method (Lieblein method) of the assessment of the depth of soil freezing. Based on data coming from observations conducted at 45 meteorological stations for over 45 years, calculations were made for the characteristic value of the zero isotherm depth with a probability of exceedance of 0.02. The obtained results illustrate the influence of climatic factors on the depth of freezing. Analyzing changes over individual decades, the pace of climate change for this phenomenon was shown, which determines the approach to this type of forecast.