The aim of the study was to evaluate the dosage effect of low-intensity pulsed ultrasound stimulation on tendon-bone healing. DESIGN: Standard partial patellectomies were performed on 120 mature New Zealand rabbits randomly assigned into the following three groups: a control group (daily mock sonication, 20 mins), a qd group (daily ultrasonication, 20 mins), and a bid group (ultrasonication twice a day, 20 mins each time). The rabbits were killed 8 or 16 wks postoperatively, and the microarchitectural, histological, and mechanical properties of the patella-patellar tendon interface were evaluated. RESULTS: Microcomputed tomography analysis showed that the bid group exhibited more new bone formation and mineralization than the other groups in the tendon-bone healing position at both 8 and 16 wks postoperatively. Histological assessments confirmed that the bid group exhibited a significantly better patella-patellar tendon interface than the other groups, as shown by the increased formation and remodeling of newly formed bone and a fibrocartilage layer. The biomechanical properties of the regenerated patella-patellar tendon interface significantly improved in the bid group. CONCLUSIONS: Low-intensity pulsed ultrasound stimulation treatment twice a day was more effective than the once-a-day treatment on tendon-bone healing.