Soil moisture and temperature conditions play an important role in plant growth. Modeling soil moisture and temperature is useful for predicting crop yields and risks. In this study, the Soil Temperature and Moisture Model (STM2) was used to predict soil moisture and temperature at several depths: 15, 30, 45, and 60 cm for soil moisture and 10, 25, and 50 cm for soil temperature. The objective of this study was to assess the prediction efficiency of STM2 according to soil depth and phenology. The STM 2uses soil texture data along with average daily weather data (maximum and minimum air temperature and precipitation) as inputs. During the 2008 and 2010 growing seasons, soil moisture and temperature data were measured using monitoring stations located in four agricultural fields in southern Quebec. These fields represent the range of soil texture diversity found in this agricultural area: gravelly, sandy, loamy, and clayey soils. The measurements were used to validate STM2 predictions. The overall performance of soil temperature prediction was better than that for soil moisture. Estimation quality decreased with increasing depth and was higher during the first and third phenological periods for soil moisture. Good performances were observed for the sandy and loamy soils, moderate for the clayey soil, and mostly weak for the gravelly soil. A sensitivity analysis was performed on STM2 data inputs. For soil moisture, bulk density, saturated hydraulic conductivity, and weather data have a great impact while for soil temperature, only weather data have an impact on model estimates. This study showed that STM2 can be used in combination with soil and climatic data sets to reliably predict surface soil moisture and temperature variations in southern Quebec.