To reconstruct the acromioclavicular (AC) joint using an adjusted closed-loop double Endobutton technique via a guiding locator that was applied using three-dimensional (3D) printing technology. At the same time, the reliability and safety of the novel double Endobutton (NDE) were tested by comparing the biomechanics of this technique with the TightRope (TR) approach. This retrospective study was conducted between January 2017 and January 2019. The Department of Anatomy at Southern Medical University obtained 18 fresh-frozen specimens (8 left and 10 right; 12 men and 6 women). First, the guiding locators were applied using 3D printing technology. After preparation of materials, specimens were divided into an NDE group, a TR group, and a normal group. In the NDE and TR groups, the navigation module was used to locate and establish the bone tunnels; after that, the NDE or TR was implanted. However, the Endobuttons were fixed while pressing the distal clavicle downwards and the length of the loop could be adjusted by changing the upper Endobutton in the NDE group while the suture button construct was tensioned and knotted after pressing down the distal clavicle in the TR. Finally, load testing in anterior-posterior (AP), superior-inferior (SI), and medial-lateral (ML) directions as well as load-to-failure testing in the SI direction were undertaken to verify whether the NDE or TR had better biomechanics. In the load testing, the displacements of the NDE and TR groups in the AP, SI, and ML direction were significantly shorter than those of the normal group (P < 0.05). In the load-to-failure testing, the ultimate load of the NDE and TR groups had significantly higher increases than the normal group (722.16 ± 92.04 vs 564.63 ± 63.05, P < 0.05; 680.20 ± 110.29 vs 564.63 ± 63.05, P < 0.05). However, there was no statistically significant difference between the two techniques for these two tests (P > 0.05). In the NDE group, four of six failures were a result of tunnel fractures of the coracoid, while two of six were due to suture breakage. In the TR, three failures were due to coracoid tunnel fractures, one was a result of a clavicle tunnel fracture, and the rest were due to suture breakage. In the normal group, half of the failures were a result of avulsion fractures of the conical ligament at the point of the coracoid process, and the other three were due to rupture of the conical ligament, fracture of the distal clavicle, and fracture of the scapular body. As for the TR technique, the stability and strength of the AC joint were better in patients who underwent reconstruction using the NDE technique than in the intact state. © 2020 The Authors. Orthopaedic Surgery published by Chinese Orthopaedic Association and John Wiley & Sons Australia, Ltd.