Two-phase flow dynamics plays an important role in understanding the kinetics of a nuclear reactor and boilers. The difficulty has been reported in obtaining the numerical solution of the two-phase flow by the two-fluid model which was described as ill-posed. Few papers have studied the difficulty under the ill-posed condition. The present paper aims at practical discussion of the property of the numerical solution of the well-defined heat-conduction equation under the above condition, comparing these solutions with the exact solution as functions of the round-off error, the mesh size and the difference scheme. The present paper can serve as an alternative in understanding the ill-posedness of the two-fluid model.