Existing experimental methods based on the measurement of crop temperature to estimate water stress have been applied for 20 years. However, the application of such techniques is limited because they are not able to totally overcome either soil interference on the measured signal or directional effects involved in temperature measurements according to sun/sensor angles configuration and crop structure. An energy balance model, based on the 3D description of plants at leaf level, is used to simulate directional cotton crop temperature variability according to crop structure and water status. The model is implemented with a bare soil compartment so that soil temperature, water balance as well heat exchanges with the crop can be computed. Once validated, this approach provides an accurate interpretation of thermal infrared information considering the directional effects involved in surface temperature measurements. This offers the opportunity of analyzing the limits of using temperature-based crop water status indices when dealing with partially covering crops. This study underlines the knowledge and tools to be further investigated in order to improve or perform such experimental techniques.