Reservoirs are strategic natural resources in particular for drinking water supply. Yet their water quality has been long influenced by anthropogenic pressures. Consequently, and in light of climate change, the frequency of harmful algal blooms events, especially cyanobacteria, has been rising and deteriorating the ecological status of these ecosystems. In this context, frequent monitoring is mandatory for tracking and understanding the dynamics of cyanobacterial blooms in time and space. Current conventional field monitoring is based on seldom sampling points which may not adequately reflect the water quality. Recent progress in remote sensing imagery makes it possible to use it for the assessment of water quality. The recent launch of Sentinel-2 satellites provides a greater coverage of water constituents with an increased spatial (10 m) and temporal resolution (5 days) suitable for monitoring small and medium-size reservoirs. However, satellite images only provide information on the water body surface layer. Three-dimensional (3D) hydrodynamic-ecological models are effective for representing chemical, physical and biological processes in the deeper layers. The objective of this work is to combine in situ measurements, satellite remote sensing and 3D modelling in order to better understand the dynamics of cyanobacterial blooms in Karaoun Reservoir, Lebanon. It is a relatively small reservoir with large water level fluctuations. The Delft3D model was calibrated, validated and applied to simulate hydrodynamics and ecological processes. Chlorophyll-a concentrations were derived from Sentinel-2 data and corrected for atmospheric interferences. Simulations of water temperature, water level and cyanobacteria biomass were compared to in situ measurements. Maps of chlorophyll-a concentrations from model simulations were compared to Sentinel-2 maps for assessing their consistency. Overall, the model represented well the hydrodynamics and ecological processes at the site. This work demonstrates the benefit of this innovative survey approach that combines remote sensing and 3D modeling techniques.