Snow melting is a complex process generally characterized by a moistening, ripening and run-off phase. Identifying the timing of these phases is crucial for river discharge, avalanches and snow contaminants monitoring. An alternative to the sporadic and punctual snow water equivalent (SWE) and liquid water content (LWC) measurements is represented by Synthetic Aperture Radar (SAR). We analyze the correlation between the multi-temporal C-band SAR backscattering and the snow melting timing. The backscattering was compared with in situ observations and snow modelling simulations for two test sites in the Alpine region. SAR sensitivity to LWC variation and snow properties enables the identification of the snow melting phases. A strong decrease of the backscattering in correspondence of an increase of LWC, e.g. during the moistening phase, is observed. Furthermore, an increase of backscattering happens in correspondence of SWE decrease, e.g. during the run-off phase. We discuss the possible mechanisms which affect SAR backscattering, related both to LWC and snow structure variations.The presented investigation could have relevant application for monitoring and predicting the snowmelt progress over large regions.