To date, no ideal heart valve prosthesis for the replacement of a diseased natural valve or for use in ventricular assist devices exists. Valves still cause thromboembolic complications originating from thrombus formations in the valve’s stagnant and recirculation zones. Optimization of valve design requires detailed flow field investigations. Usually, the regions that are more prone to thrombus formation can be estimated using a dye washout experiment. This successful experimental method was simulated using numerical methods. The proposed method was applied to three standard mechanical heart valves—Björk-Shiley, St-Jude, and Starr-Edwards valve. The dye washout was characterized by a time course of the gray value averaged over a defined region of interest. Finally, these curves were quantified by a half dye time (HDT), which characterizes the blood residence time. The HDT in the best valve, the Starr-Edwards valve, was 0.0747 s. The HDT in the worst valve, the Björk-Shiley, was 0.0942 s. The analysis of the hemodynamic valve parameters (pressure drop, velocity magnitudes and turbulence) revealed that the best valve is the St-Jude valve. The Starr-Edwards valve displayed the worst hemodynamic parameters. This study shows that the proposed numerical method of dye washout visualization can be used as an additional tool for the flow characterization.