Ligament reconstructive surgeries demand tunnel creation using an over-drilling technique, though this technique has some problems such as metallic particle liberation or difficulties in tunnel creation other than circular cross-section. Recently, a new ultrasound (US) device for bone excavation to overcome these problems was developed. This study aimed to compare the tendon-bone healing in tunnels created using the new US device to that created using the conventional drill in a rabbit model. A total of 72 rabbits underwent a reconstruction for the anterior half of the medial collateral ligament (MCL) using a half of the patellar tendon. For the femoral tunnel creation, a new US device was used in 36 rabbits (US group), while a conventional metallic drill was used for the remaining 36 rabbits (DR group). At 4, 8, and 12 weeks postoperatively, biomechanical (n = 10) and histological (n = 2) evaluations were performed. The ultimate failure load was almost equivalent between the US and DR groups at each period (US/DR; 4 weeks, 50.0 ± 12.8 N/43.4 ± 18.9 N, p = 0.62; 8 weeks, 78.6 ± 11.5 N/77.3 ± 29.9 N, p = 0.92; and 12 weeks: 98.9 ± 33.5 N/102.2 ± 38.3 N, p = 0.80). Pull-out failure from the femoral tunnel was only observed in two rabbits in the US group and one rabbit in the DR group at 4 weeks postoperatively. At 8 and 12 weeks, all specimens had a mid-substance tear. The collagen fiber continuity between tendon and bone occurred 8 weeks postoperatively in both groups and no histological difference was recognized throughout the evaluation period. The tunnels created using the new US device and the conventional drill had equivalent biomechanical and histological features in tendon-bone healing. The bone excavation technology by the new US device may be applicable in ligament reconstructive surgeries. Copyright © 2020 The Japanese Orthopaedic Association. Published by Elsevier B.V. All rights reserved.