Canal seepage is of great concern because of the water shortage and efficient use of irrigation water resource in many irrigation districts. To investigate the infiltration loss, heat was used as a natural tracer to characterize the canal seepage during irrigation channels conveyance. The field experiment was carried out in a typical canal located in Hetao Irrigation District. Time-domain reflectometer and thermal sensors were used to monitor the soil moisture content and temperature around the typical cannel. The tracer experiment indicated that temperature changes in the shallow sediments characterized canal seepage properties. Based on the interactions between groundwater and surface water temperature, temperature changes can derive the river flow even without water level data. Two-dimensional water and heat model, VS2D, was applied to simulate the dynamic process of water and heat in the saturated and unsaturated soils. And the soil hydraulic parameters, including saturated water content and hydraulic conductivity were inversed by the VS2D model. The hydraulic conductivity was 0.025m/d for the silt loam and 0.6m/d for the loamy sand soil in the experiment area, respectively. The simulation results were verified based on the results of the observation experiment. The values of root mean square error (RMSE) and mean relative error (MRE) between simulated and measured temperatures were 1.6℃ and 2.5%, respectively, indicating an accurate simulation. The relative mean error between predicted accumulative infiltration amount and actual value was 2.4% and the accuracy could meet the requirement. The results suggested that heat as a tracer was reliable for monitoring of water transport during canal seepage. Heat as a tracer provided a convenient and accurate way to monitor the infiltration loss during irrigation channels conveyance, which would be beneficial to irrigation management in the arid irrigation district in North China.