Abstract: Due to the limitation of the recognition scope of Leap Motion（LM） and its image recognition algorithm, finger occlusion phenomenon may occur when finger plane is parallel to the detection direction of LM. It further affects accuracy and interaction experience of gesture recognition. In order to solve the problem of finger occlusion to improve the interaction experience, we propose a method to set two LMs at different angles. First, we set the two LMs vertically at the predetermined positions. Then we conduct matrix transformation for the detected data of above LMs. Data fusion process is further performed to obtain gesture types of fingers movements. User studies are conducted to demonstrate the useful-ness and effectiveness of our method in reducing finger occlusion and improving usability.