The accelerated changes in land use and human society during the last decades, and the consequent anthropogenic stresses applied to water resources, have led eutrophication to be one of the major freshwater environmental problems. At the same time, the unprecedented rate of climate warming contributed to an increase in frequency, intensity, and duration of cyanobacteria blooms. For these reasons, an ever increasing interest is given by the scientific community to ecological modeling in order to give water quality managers reliable projections for decision making. Numerous ecological models based on various approaches arose in the last decades. However, modeling the dynamics of cyanobacteria in inland water bodies remains a challenging task. This is mostly due to the generalized lack of suitable data, crucial to calibrate and validate model results as well as to keep improving the formulations of models themselves. We analyse here the performances of two different approaches in the modeling of cyanobacteria dynamics by confronting model results to field data. The first ecological model is Aquatic EcoDynamics (AED), a customizable modular model library based on the conceptualization of biogeochemical cycles and limitations in growth rates for numerous phytoplankton functional types in the classic form of ordinary differential equations [1]. The second one is the BLOOM module of the Delft3D modeling suite [2]. It also involves nutrients cycles and growth rates in its formulation, but relies on a competition principle that is able to choose, at every time step and each grid point, the best adapted phytoplankton functional type consistently with the available resources and the existing biomass level. Both ecological models are coupled to three-dimensional hydrodynamic models, Telemac3D and Delft3D-Flow respectively. The study site, Lake Champs-sur-Marne, is a small and shallow urban lake (area 0.12 km², average depth 2.3 m), located in the metropolitan area of Paris and used for bathing and sailing. It suffers from repeated severe blooms of cyanobacteria from spring to fall and since 2015 it has been more regularly monitored with high frequency measurements and bi-weekly surveys. Simulations of both ecological models have been carried out over three consecutive summers during critical blooming periods. The modeling results on each blooming period have been compared to the collected data from Lake Champs in order to evaluate the reliability of their different formulations in the context of future projections under climate change scenarios.