Even if the interest in observing soil water content is high from decades, there is no one technology that the majority of soil scientists can consider to be definitive. In particular TDR methodology, a by product of a technology developed to analyze transmission lines, has been for years considered to be based on a theory complete enough for applications in soil science. This made most scientists to focus on interaction with soil (antenna design) or simplifying the systems to reduce costs. Even if probably the major problem of EM techniques stay in the representation of the structure of the medium, much has been done in this years on strip lines theory and in particular on propagation in non-uniform media, which can make TDR worthwhile to be reconsidered in deep. Recently classic TDR approach have been investigated to use the reflected signal from the probe to obtain soil moisture distribution by means on inversion technique. In the present investigation classic TDR technology, namely that based on Tektronics 1502, has been re-analysed with a theoretical-methodological-experimental approach. Theory of Telegraph equation for non-uniform case has been revisited and an inversion technique has been derived to get from each reflected data sequence a soil moisture distribution. Both concentrated and distributed parameter of each part making the transmission line has been tested in laboratory. The probe has been designed with the objective of solving most sensitive problems of laboratory and field soil measurement protocols, common in TDR practice, and overall minimization of disturbance of soil in field insertions. The technique has been verified directly on soil samples, and the profile reconstructed from measurement of water content from a single TDR signal on most of their length. Test have been made on sand and clayey soil (disturbed and un-disturbed), allowing to reconstruct values of water content with an error less than 2% on a 2 cm layer of soil.