The article reviews the use of optical remotely sensed time series data for land cover classification, emphasizing the importance of accurate land cover information for science, monitoring, and policy-making. It highlights the challenges and opportunities associated with generating and validating annual, large-area land cover products using Earth Observation (EO) data, particularly from the Landsat archive and the upcoming Sentinel-2 data stream. The review discusses the benefits of time series data, such as improved accuracy and the ability to capture class stability and logical transitions. It also explores various classification methods, including unsupervised and supervised approaches, and the integration of novel inputs like ecological succession knowledge and multi-scale, multi-sensor data fusion. The article underscores the need for robust techniques to handle temporal data and improve the temporal consistency of land cover maps, while also addressing the limitations and potential improvements in current methods.The article reviews the use of optical remotely sensed time series data for land cover classification, emphasizing the importance of accurate land cover information for science, monitoring, and policy-making. It highlights the challenges and opportunities associated with generating and validating annual, large-area land cover products using Earth Observation (EO) data, particularly from the Landsat archive and the upcoming Sentinel-2 data stream. The review discusses the benefits of time series data, such as improved accuracy and the ability to capture class stability and logical transitions. It also explores various classification methods, including unsupervised and supervised approaches, and the integration of novel inputs like ecological succession knowledge and multi-scale, multi-sensor data fusion. The article underscores the need for robust techniques to handle temporal data and improve the temporal consistency of land cover maps, while also addressing the limitations and potential improvements in current methods.