Although guided high-mechanical-index (MI) impulses from a diagnostic ultrasound transducer have been used in preclinical studies to dissolve coronary arterial and microvascular thrombi in the presence of intravenously infused microbubbles, it is possible that pulse durations (PDs) longer than that used for diagnostic imaging may further improve the effectiveness of this approach. By use of an established in vitro model flow system, a total of 90 occlusive porcine arterial thrombi (thrombus age: 3-4 h) within a vascular mimicking system were randomized to 10-min treatments with two different PDs (5 and 20 μs) using a Philips S5-1 transducer (1.6-MHz center frequency) at a range of MIs (from 0.2 to 1.4). All impulses were delivered in an intermittent fashion to permit microbubble replenishment within the thrombosed vessel. Diluted lipid-encapsulated microbubbles (0.5% Definity) were infused during the entire treatment period. A tissue-mimicking phantom 5 cm thick was placed between the transducer and thrombosed vessel to mimic transthoracic attenuation. Two 20-MHz passive cavitation detection systems were placed confocal to the insonified vessel to assess for inertial cavitational activity. Percentage thrombus dissolution was calculated by weighing the thrombi before and after each treatment. Percentage thrombus dissolution was significantly higher with a 20-μs PD already at the 0.2 and 0.4 MI therapeutic impulses (54 ± 12% vs. 33 ± 17% and 54 ± 22% vs. 34 ± 17%, p < 0.05 compared with the 5-μs PD group, respectively), and where passive cavitation detection systems detected only low intensities of inertial cavitation. At higher MI settings and 20-μs PDs, percentage thrombus dissolution decreased most likely from high-intensity cavitation shielding of the thrombus. Slightly prolonging the PD on a diagnostic transducer improves the degree of sonothrombolysis that can be achieved without fibrinolytic agents at a lower mechanical index.