Long-term trends in soil temperature are important, but rarely reported, as a sensitive climate indicator and its impact on biological, chemical, and physical processes that occur in the soil. In this study, the simple linear regression and nonparametric Mann-Kendall (MK) trend test were employed to identify trends in monthly air and soil temperature anomalies (ATA and STA, respectively) time series at 5-100 cm depths of a soil profile, based on long-term (1987-2016) observation data of meteorological station of Bajgah Agricultural Experimental station, located in a semi-arid region in Iran. The results indicated a significant warming trend in ATA for February-October, with the highest and lowest rate of warming in March (0.103 degrees C year(-1)) and December (-0.001 degrees C year(-1)), respectively. This study further revealed a general and strong increase in monthly means of STA over the studied period. The rates of soil temperature warming trend were 2-3 times greater than those of air temperature. The greatest soil warming increases with the rates of 0.20 and 0.19 degrees C year(-1) were, respectively, observed at 5 and 10 cm soil depths in July. The applied linear trend indicated that the STA at 10 and 30 cm depths have warmed during May-October. In addition, the MK test revealed a significant trend in STA at 10 cm depth in March. Such significant warming trends in subsurface soil depths including 50 and 100 cm were observed for June-December during the whole period.