We developed and successfully tested a differential scanning calorimetry (DSC) method to estimate the liquidus temperature (T(L)) of good glass-forming systems, i.e., that are reluctant to crystallize. The method was first tested for several Li(2)O-B(2)O(3) glasses. The onset, peak, and endpoint temperatures of DSC melting peaks were measured and compared with the liquidus in the phase equilibrium diagram. DSC runs were carried out at different heating rates and T(L) at equilibrium was estimated by extrapolation to 0 degrees C/min. For glasses that do not crystallize during a typical DSC run, a previous heat treatment was necessary to induce crystallization. The liquidus of three multicomponent glasses were then obtained by DSC and compared with the T(L) obtained by a direct technique using optical microscopy. The endpoint of the DSC melting peaks measured at different rates and extrapolated to 0 degrees C/min was the best estimate for the liquidus of such glasses. Most results differ by not more than 10 degrees C. Because of the small amount of glass needed, the possibility of instrumental detection, simultaneous crystallization pretreatment of many different compositions, and the speed of the DSC analysis, the proposed technique may be a valuable option to estimate the liquidus of simple or complex multicomponent glasses.