Based on the model-based virtual sensing approach, previously proposed by the authors, two melt flow virtual sensors are developed. The virtual sensors estimate the melt front position during the filling process by using easily obtained machine variables, specifically, the injection hydraulic pressure and the screw position and nozzle pressure. Both virtual sensors were theoretically analyzed and experimentally evaluated on a commercial injection molding machine. The theoretical analysis provides step-by-step design guidelines for the virtual sensors. Performance evaluation with a twin thin-plate mold validates the feasibility of the proposed model-based melt-flow virtual sensors.