It is of critical importance to acquire tree transpiration characters through sap flow methodology to understand tree water physiology, forest ecology and ecosystem water exchange. Tri-probe heat pulse sensors, which are widely utilized in soil thermal parameters and soil evaporation measurement, were applied to implement Salix matsudana sap flow density (Vs) measurements via heat-ratio method (HRM), T-Max method (T-Max) and single-probe heat pulse probe (SHPP) method, and comparative analysis was conducted with additional Grainer's thermal diffusion probes (TDP) measured results. The results showed that, it took about five weeks to reach a stable measurement stage after TPHP installation, Vs measured with three methods in the early stage after installation was 135%-220% higher than Vs in the stable measurement stage, and Vs estimated via HRM, T-Max and SHPP methods were significantly linearly correlated with Vs estimated via TDP method, with R2 of 0.93, 0.73 and 0.91, respectively, and R2 for Vs measured by SHPP and HRM reached 0.94. HRM had relatively higher precision in measuring low rates and reverse sap flow. SHPP method seemed to be very promising to measure sap flow for configuration simplicity and high measuring accuracy, whereas it couldn' t distinguish directions of flow. T-Max method had relatively higher error in sap flow measurement, and it couldn' t measure sap flow below 5 cm3 · cm(-2) · h(-1), thus this method could not be used alone, however it could measure thermal diffusivity for calculating sap flow when other methods were imposed. It was recommended to choose a proper method or a combination of several methods to measure stem sap flow, based on specific research purpose.