Objective: Rotary blood pumps (RBP) are long-term mechanical circulatory support devices that support a failing heart by pumping blood. Ventricular collapse and suction, which can lead to myocardial damage and arrhythmias, is a significant risk factor during RBP support. The RBP also needs to maintain pump flow to match perfusion demand over a wide range of physiologic conditions. We have developed a novel sensorless control algorithm to maintain physiologic perfusion while avoiding ventricular suction, using a suction index (SI) extracted from the intrinsic pump speed measurements. Methods: The objective of the proposed control algorithm is to maintain an SI setpoint. Using nonlinear mathematical models of a human circulatory system and a RBP, efficacy and robustness of the proposed algorithm with 2% RPM measurement were tested in-silico by comparing it to differential pump speed control, differential pump pressure control, constant speed control, and mean aortic pressure control during (1) rest and exercise conditions, (2) a rapid eight-fold increase in pulmonary vascular resistance for rest and exercise, and (3) transition from exercise to rest. Results: The proposed control algorithm provided physiologic perfusion while simultaneously preventing ventricular suction for all test conditions. The performance of the proposed control algorithm was superior to other tested control strategies in avoiding suction. Conclusion: Maintaining a reference SI effectively provided physiologic perfusion and prevented ventricular suction. Significance: The proposed SI control approach can meet physiologic circulatory demand and avoid suction in RBP without requiring the use of unreliable pressure or flow sensors or a pump model.