Developing sustainable crop varieties for climate change, drought, and heat waves is an immediate need for the agricultural community. Breeding scientists are actively developing better crop varieties; however, assessing the performance of thousands of genotypes, called plant phenotyping, presents a significant logistical challenge. Unmanned Aerial Vehicle (UAV) -based Hyperspectral imaging (HSI) is a potential technology to address this challenge, which acquires rich spectral information of objects in hundreds of spectral channels with large fields of coverage. However, very few crop HSI datasets are available for public use in the literature. In this study, we leveraged UAV-based HSI covering the spectrum range of 400–1000 nm to create a comprehensive dataset of pearl millet canopy water stress. Utilizing state-of-the-art machine learning (ML) benchmarks, we conducted classification experiments to showcase the dataset’s effectiveness in characterizing progressive water stress in pearl millet canopy. Moreover, we emphasize our commitment to making this dataset publicly available, recognizing its potential to advance research activities in the agriculture domain using UAV-based sensing.