Fuel moisture content (FMC) of live vegetation is an important factor for wildfire ignition and propagation. Therefore, for wildfire danger assessment, it is inevitable to consider this parameter in models regarding the prediction of the ignition probability and spreading behaviour of wildfires. A current prototype of an integrated forest fire danger assessment system in Austria is introducing FMC as a functional derivative of meteorological conditions,which was calibrated for Canadian pine trees. To improve wildfire danger assessment in Austria, an updated FMC dataset that considers the local environmental and vegetational conditions is required. Observations from optical remote sensing instruments are sensitive to FMC in the infrared domain and thus, offering the possibility to estimate this quantity on a broader scale.An established global dataset for FMC is derived from the Moderate-Resolution Imaging Spectroradiometer (MODIS), which delivers temporally frequent reflectance data with a spatial resolution of 500 m. Modelling FMC with this pixel size might be too coarse when considering narrow valleys or mountain peaks. In this study, satellite observations from theSentinel-2 mission were used to calculate a new FMC dataset for the area of Styria with a high spatial resolution of 20 m. The aim was to develop a workflow that should further allow the FMC estimation for the entire Austrian country, which can support the domestic wildfire danger assessment. This was done by applying a pretrained machine learning model to Styrian Sentinel-2 data. This model was trained on Australian Sentinel-2 observations that were related to the MODIS derived FMC values. The validation of this approach with MODIS FMC data of Styria led to rather low correlations (ρ between 0 and 0.16) between the two datasets.A main reason for the low correlation results was identified as the relevant climatic and vegetational differences between Australia and the study area of Styria. Another reason was the different origin of the data. While the Sentinel-2 reflectances of Australia were subjected to a preprocessing algorithm, the Styrian data was originally taken from the Google Earth Engine tool. To develop a valid FMC dataset for Austria based on optical remote sensing data and machine learning algorithms, the training of a new model is recommended for further research work. The inclusion of a data layer that contains locally optimized FMC estimations with high spatial resolution remains a desirable goal for the Austrian wildfire danger assessment.