Sucrose and its supersaturated or supercooled solutions are of the utmost importance in areas as different as food and pharmaceutical industries or cryopreservation of biological systems. In the supercooled "state", such amorphous solutions are fragile systems, presenting extremely high viscosity with decreased mobility of molecules, impairing experimental determinations at low temperatures (below critical temperature, T-c). This work proposes the random walk (RW) approach to study relaxation behavior in the supercooled melt, using only viscosity data at temperatures above T-c. For comparison purposes, the Vogel-Fulcher-Tamman (VFT) model was fitted to th data. The random walk approach is based on the distribution of molecular energies, thus having a theoretical support not found in the VFT model. The RW estimated parameters can be correlated with system characteristics, showing their dependence on water content or sucrose concentration. RW and VFT model description was evaluated from the mathematical and physical points of view, but the random walk model led to more consistent regression results and parameter estimates with improved precision.