In this study, soil temperatures at different depths in Turkey's different regions were investigated theoretically. Soil temperature data are critical for different research interests such as ecology, biology, technique processes, forestry, agriculture, energy, food sector, ground heat exchanger applications, thermal energy storage applications, and so forth. This investigation gives information related to the prediction of soil temperature's dependence with depth and time especially for shallow geothermal applications. Soil temperature values depend on a great deal of varied parameters such as thermal conductivity, short term climatic conditions and moisture content. The main issue is that despite these temperatures are extremely important values, they can not be obtained in a short time. Due to this reason, we study a mathematical model related to the prediction of soil temperature. Within this context, 81 cities and their approximately 300.000 data, both, monthly air and soil temperatures between 1960 and 2015 were studied and finally seven regions in Turkey were investigated and final average soil temperature values were achieved. Measured data taken from the Izmir State Meteorological Station, and predicted soil temperatures at depths of 5 cm, 10 cm, 20 cm, 50 cm, and 100 cm were analyzed for each region in Turkey according to data obtained fifty years ago. Finally, at depths of 5 cm, 10 cm, 20 cm, 50 cm and 100 cm, the maximum average percentage errors in Turkey were 16%, 14.8%, 13.5%, 14.4%, 13.9% respectively. In conclusion, we evaluate the relationship between ambient air temperatures and soil temperatures in terms of depths from 5 to 3000 cm.