Various studies have demonstrated the potential of thermal infrared brightness temperature (TIR TB) for monitoring surface exchanges of water and energy. This study focuses on the contribution of TIR TB data for Land Surface Model (LSM) calibration. A numerical representation of the Soil-Vegetation-Atmosphere (SVA) transfers (SVAT model), named SEtHyS, was used. A calibration methodology of the model based uniquely on the optimisation of TIR TB diurnal cycle features has been developed and applied, in an assimilation context, to the full vegetation period of a wheat crop. The results illustrate the advantages of such a methodology for the monitoring of environmental conditions simulated with the SVAT model, such as the root zone soil moisture. The impact of observation and simulation errors on TIR TB was analysed and quantified in controlled numerical experiments. The results demonstrate the advantages of using relative temperature characteristics, instead of temperature values themselves, to minimise the impact of noise.