The present study aims at developing a specific algorithm to automatize the calibrated average-gradient (CAG) heat pulse method and estimating low and reverse sap flow in roots, termed as CAGALG. The innovative algorithm is based on an automatic data processing to make easier the complexity in interpreting data from sap flow sensors, without losing the reliability on low and even reverse sap flows quantification. The CAGALG algorithm resorts to two mathematical tools to indicate reverse sap flow occurrence and clustering the roots per reverse sap flow patterns. Results showed a good correlation between the average temperature gradient (ΔTavg) and the corrected heat pulse velocity (VC) that was confirmed by the simulation study. The simulation provided a model aiming to estimate the thermal diffusivity, which has been shown to be below the default values usually attributed. Quantitative and temporal reverse sap flow occurrences were identified along the two years in which presented the results from the CAGALG in accordance with the traditional CAG.