Soil heat and moisture processes are interconnected, especially during low temperatures. To examine the interaction between soil temperature and moisture under freeze-thaw cycles, a physical process-based model (CoupModel) coupled with uncertainty analysis was applied to 3-year measurements under seasonal frost conditions from a site in the black soil belt of northeast China. The uncertainty in parameters and measurements was described by general likelihood uncertainty estimation (GLUE). To identify the degree of linkage between soil temperature and moisture, three criteria were applied to them separately or together. The most sensitive parameters among 26 site-specific parameters were closely related to soil heat, soil evaporation and freeze-thaw processes. Soil temperature was simulated with less uncertainty than soil moisture. Soil temperature measurements had the potential to improve model performance for soil water content, whereas soil moisture measurements demonstrated a trade-off effect when finding a model with good performance for both temperature and moisture. During winter conditions the uncertainty ranges of soil temperature were most pronounced, probably because of the greater complexity of soil properties during the freeze-thaw process and the uncertainty caused by snow properties. The largest uncertainty ranges of both soil water content and soil water storage were found mainly in the deep soil layers. The simulated surface heat fluxes are an important output of the model and it is of great value to compare them with the results from regional climate models and micrometeorological measurements.