The project F-Predict is dedicated to creating a prediction model which can predict fertilizer prices. The experiences in financial forecasting which were gained in the project smart farming are used. Different machine learning models will be applied. The focus is on signal selection. The evaluation uses historical, test and real data.