Abstract: In the task of high-precision radar-based aerial target recognition, single-view inverse synthetic aperture radar (ISAR) is susceptible to azimuth variations, partial occlusions, and noise interference. These factors often hinder the representation of a target′s complete structure, leading to reduced recognition accuracy. To address this issue, this paper proposes an aerial target recognition method based on multi-view ISAR image fusion. The method employs a shared feature extraction network to extract semantic features from multi-view ISAR images and utilizes a Transformer encoder to model global dependencies across different viewpoints. Additionally, a dynamic gating fusion module is introduced to adaptively weight and fuse discriminative features from various ISAR viewpoints. L1 regularization is applied to constrain the viewpoint weights of ISAR images, effectively enhancing the network′s ability to suppress redundant views. The proposed method is validated through electromagnetic scattering simulations on six types of aerial aircraft targets. Experimental results demonstrate that, compared to traditional single-view methods, the proposed approach improves recognition accuracy by approximately 15 % and 20 % under high and low signal-to-noise ratio (SNR) conditions, respectively.