The aim of this thesis project is to investigate the temporal changes of vegetation in Europe using statistical analysis and machine learning. A methodology is used consisting of five main steps. The first is data pre-processing, which is used to reduce the noise within the data set. This is followed by classification, which divides the data into its various land cover types. Next, the time series are decomposed into their trend and seasonal components. The fourth step is to forecast these time series components, and finally the generated trend and seasonal components are analysed. Annual land cover results for the entire domain have been produced, with an over-all classification accuracy exceeding 80%. Furthermore, a statistically significant slow degree of greening was determined for the majority of the domain over the twenty year time span and the growing season shows lengthening for the areas that transitioned from browning to greening.