We developed a comprehensive, effective model to account for resistive losses along the feeding line (commonly termed cable losses) in TDR measurements of electrical conductivity. The feeding line including cable, connectors, and multiplexers is modeled as a generic two port network, and its behavior at low frequency analyzed using standard network theory. We found that the line is characterized by two parameters, effectively corresponding to longitudinal and transverse losses, and which can be easily obtained through calibration using the line with probe in air and short-circuited. The model does not require prior knowledge of the cell constant, and was verified using measurements of sample resistance with multiple different feeding line configurations. Results confirm the efficacy of the model, and show that the commonly-used two-resistors model works only for a conductivity range that is greater than the range of interest in many soils applications. The incident reflection voltage Vi should be measured immediately following the incident step. The new model can be used to compensate for errors resulting from use of other cable loss models, but those due to incorrect measurement of the incident and final voltage cannot be corrected, unless these values are known.