The goal is models for translating satellite data from the time of supplementary fertilisation directly to nitrogen (N) recommendation maps. The crops are winter wheat (Triticum aestivum L.) and malting barley (Hordeum vulgare L.). A new multispectral camera for drones, with nine bands of the same specifications as Sentinel-2 in the 400-900 nm spectral range, will be used to collect reflectance data in Swedish trial series L3-2299 and L3-2302. Robust models for economically optimal N rates will be developed and evaluated. If successful, models can be implemented in satellite-based decision support systems for precision agriculture, such as CropSAT. Key results will be implemented in a ‘Focus on Nutrients’ (national programme to improve nutrient use efficiency) module on precision agriculture. Potential benefits of the models can be large, by enabling wide application of more accurate N fertilisation. The project spans the knowledge chain from needs-driven research to evaluation.