In the past, large-scale terrestrial vegetation dynamics have been used to evaluate the efficiency of national ecological restoration programmes in northern China. However, complete documentation and timely updates of vegetation dynamics in the Ecological Restoration Programmes Regions (ERPRs) of northern China are rare. Furthermore, the main challenges in analysing large-scale vegetation changes in ERPRs are so far not entirely clear. We chose four ERPR study areas (the Beijing-Tianjin Sand Source Region (BTSSR), the Loess Plateau (LP), and the north-east (NE), and the north-west (NW) of the Three-North Shelterbelt Forest Project (TNSFP)), avoiding the overlap of administrative regions. We used long-term Normalised Different Vegetation Index (NDVI) data to investigate the trends of vegetation activities in these four ERPRs of northern China between 1982 and 2011. A persistent and widespread increase of the growing season NDVI was found in this time span, with 58.54% increase in NE, 89.84% in LP, 54.59% in BTSSR, and 76.13% in NW. The observed increase of NDVI indicates the success of Chinese restoration programmes, which have helped to improve ecological conditions after 2000. However, a comparison of long-term vegetation changes between before and after implementation of restoration programmes is insufficient for assessing the programmes' efficiency. Large-scale Dynamic Global Vegetation Models (DGVMs) are promising tools for future evaluation of the benefits of ecological restoration programmes.