Image-based ship target analysis is an important task in the field of ship monitoring. Previous studies have achieved remarkable results in ship detection and recognition tasks. However, these related studies mainly rely on unimodal datasets, and there is still no publicly available ship individual re-identification dataset released, which restricts the research in the field of cross-modal individual re-identification of ship targets. To address this issue, we have constructed the first cross-modal ship re-identification dataset, CMShipReID. This dataset contains data from three modalities, namely visible light, near-infrared, and thermal infrared, which are collected by drones. It covers 10 categories, approximately 138 individual ships, and 8,337 images, thus providing data support for the research on cross-modal individual re-identification of ships. We have tested the mainstream re-identification algorithms as the performance benchmark for this dataset, which can serve as a fundamental reference for relevant scholars.