With this dataset, we provide raster data containing 20 predicted plant functional traits based on EnMAP imagery from 2022 to 2024 in the Arctic biome. The traits were predicted based on a convolutional neural network created by Cherif et al. (2023) and adapted specifically to EnMAP data by Mederer et al. (2025). We used all available EnMAP tiles during the vegetation periods from 2022 until 2024 that were obtained in the Arctic region. However, due to the targeted observation scheme of the EnMAP mission, the tiles mainly cover regions on the North American continent and Greenland. Each raster file (in .tif- format) contains 20 layers, one layer per predicted plant trait. A water mask based on NDVI < 0.1 has been applied to all scenes. For further details on model architecture the user is referred to the above mentioned publications. The data was collected in order to assess ecosystem functional diversity and classify land cover types in a trait-based framework across the Arctic biome. It can be further used for mapping purposes, including the distribution of plant functional types. The data can be paired with field spectroscopy for upscaling approaches as well as for fusion with imagery from other sensor types. For now, this dataset can be considered a preliminary version. Meanwhile, the model for trait prediction is being under further development to address the specific requirements of Arctic vegetation for trait predictions from hyperspectral data.