<p>Vehicle re-identification (Re-ID) has become a challenging retrieval task due to the high inter-class similarity and low intra-class similarity among vehicles. To address this challenge, the self-attention mechanism has been extensively studied and applied, demonstrating its effectiveness in capturing long-range dependencies in vehicle Re-ID. Traditional spatial self-attention and channel self-attention assign different weights to each node (position/channel) based on pairwise dependencies at a global scale to model long-term dependencies, but this approach is not only computationally complex but also unable to fully mine refined features. In this paper, we propose a vehicle Re-ID network design based on a multi-axis compression fusion (MCF) attention mechanism. The MCF attention mechanism preserves feature information on different axes through compression operations while maintaining high computational efficiency. It utilizes single-axis self-attention calculations to update the weights and strengthens the regions of common interest across multiple axes by fusing information from multiple axes, thereby enhancing the effect of attention learning. On the basis of this mechanism, we propose a multi-axis compression fusion network (MCF-Net), which combines the spatial multi-axis compression fusion (S-MCF) module and the channel multi-axis compression fusion (C-MCF) module, and uses a rigid partitioning strategy to capture both global and fine-grained features. Experiments show that MCF-Net achieves state-of-the-art performance on the vehicle Re-ID datasets VeRi-776 and VehicleID.</p>