The integrated and sustainable management of forest ecosystems provides substantial environmental, societal, economical and other benefits and contributes towards the provision of a wide range of products and services for a sustainable exploitation and use of forest resources. Towards this aim, it is necessary to continuously monitor the forest ecosystems for achieving an increase of their primary production, while, at the same time, preserving their biodiversity and their sustainable use. Modern geospatial technologies, such as satellite remote sensing, when combined with the rapid advances in big data analysis and machine learning fields, provide powerful tools for the development of reliable and (often) low cost systems of monitoring land and non-land ecosystems through new automated processes and tools. In this paper, a proposed automated workflow is presented, which has been developed within the ARTEMIS Greek research project, aiming at: a) preprocessing of time series of geospatial data (mainly free satellite data from Sentinel and Landsat satellites), b) the extraction of vegetation and forest health indices, and c) the identification of basic vegetation/forest categories and estimation of classification maps for the specific Area Of Interest (AOI). Particular emphasis will be given to the use of modern powerful machine learning tools, such as recurrent neural networks (such as LSTM, which is known to exhibit excellent performance in temporal classification problems). After the completion and validation of similar workflows for satellite image analysis, the final system will be supported by a WebGIS platform (either open source or proprietary, such as ERDAS Apollo of Hexagon Geospatial), which will provide new products and services to the final users of ARTEMIS project. For these reasons, these workflows will be thoroughly evaluated, based on performance versus cost criteria, so that those that will be integrated to the final digital platform of the project will be determined, offering new on-demand data analysis capabilities to its final users.