Terrestrial vegetation plays a key role in land surface-atmosphere interactions as the primary link for moisture (evapotranspiration) and energy (latent) exchange. Consequently, the regional to global effects of historical and expected land use/cover changes (LUCCs) due to both natural and human factors remains a key subject in environmental research. In the East Africa region, LUCC is a key subject as approximately 30% of the region has been converted to cropland or urban areas in the recent past. The recent effects of droughts on vegetation dynamics in the region have also put a spotlight on the potential effects of climate change on ecosystems in the region. This study focuses on the assessment of the vegetation dynamics and effects of land cover changes on the water and energy balance in the East Africa. To address the complexity of the research, an integrated approach was adopted linking climate observations, remote sensing data on vegetation phenology and land surface fluxes, and a land surface model (Variable Infiltration Capacity (VIC) model). Satellite-based Leaf Area Index (LAI) observations over the 1982 to 2011 period showed a significant increasing trend in about 25% of the region while decreasing trend was recorded in 4% of the region. The spatial-temporal vegetation dynamics showed varied relationship with both climate anomalies and human activities. Long-term vegetation dynamics derived from MODIS datasets were used to parameterize VIC model in order to assess the regional sensitivity of the surface water and energy fluxes to varying seasonal and interannual vegetation changes. The results were compared to globally tested datasets on Evapotranspiration (ET), Latent heat and Sensible heat based on FLUXCOM global datasets on surface energy balance as well as MODIS evapotranspiration data. The adoption of vegetation parameter values varying spatially and in time indicated a more realistic representation of the water and energy fluxes in the region. Across the region, the variations in vegetation parameters is found to impose higher standard deviation on mean annual sensible heat (0.3 Wm-2) than on latent heat (0.03 Wm-2) but generally has small effect on ground heat (0.02 Wm-2). However, the percentage of positive variance imposed on ground heat by varying vegetation parameters across the region is higher (33%) compared to 13% and 14% for ET/latent heat and sensible heat, respectively. Using the newly configured VIC model for east Africa, effects of LUCCs were simulated mainly focusing on the potential effects of increased tree cover through afforestation or/and reforestation efforts on regional LAI, FVC and albedo. The land cover changes were simulated using the Dynamic Global Vegetation Model with managed Land, (LPJmL4) with parameters adapted for the east Africa region. This approach allowed for simulation of increased spread of tree cover in bare areas and areas covered by shrubs and grass, while accounting for habitat suitability for growth of different tree vegetation types. The simulated increase in tree cover across the region showed potential increase in mean LAI and fractional vegetation cover by 1.5 m2/m2 and 0.16 respectively, while albedo decreased by 0.02 averaged across the region. The simulated changes in the water balance due to increased tree cover were mainly linked to the increased transpiration and canopy evaporation as well as decrease in soil evaporation. Particularly, increased tree cover led to increased latent and sensible heat while land surface temperature and canopy temperature decreased by -1.0°C and -1.9°C respectively. Based on the findings of the study, increasing forest cover in the region will play a pivotal role in climate change adaptation by inducing a highly needed local cooling effect. However, there is still need for extensive analysis of the potential impacts of natural and human-induced vegetation dynamics in East Africa at a much smaller scale. This will particularly require reduced uncertainties in the region’s assessment studies particularly through better parameterization of small-scale variations in vegetation as well improved regional constraining of land surface models using multiple datasets.