Understanding the spatial and temporal dynamics of cyanobacterial blooms requires mechanistic models that capture interactions between physiological traits and physical lake processes. We developed a three-dimensional (3D) individual-based model (IBM) to simulate the growth, loss, and buoyancy-driven movement of Dolichospermum sp., coupled with a 3D hydrodynamic model that incorporates advection and dispersion processes. The modelling system was applied to Peter Lake (Michigan, USA) during a whole-lake nutrient enrichment experiment in the summer of 2015. To represent intra-specific variability, the IBM considered a range of physiological trait values—specifically, maximum growth rates (0.4–1.1 day-1) and floating velocities (0.2–10 m day-1)—allowing for a more realistic representation of population-level responses under dynamic environmental conditions. The model also incorporated antecedent environmental conditions to account for non-photochemical quenching—an adaptive physiological trait—thereby improving biomass predictions validated against observed cyanobacterial cell counts. Model results showed that thermal stratification and mixing significantly influenced bloom dynamics. Stratification allowed Dolichospermum filaments to float toward the surface and form blooms under optimal growth conditions, while subsequent deepening of the mixed layer redistributed filaments and increased light limitation, leading to bloom collapse. The IBM demonstrated that floating velocity plays a critical role in shaping the environmental conditions experienced by cyanobacteria. Filaments with higher floating velocities accumulated near the water surface, where they were exposed to higher light intensities and warmer temperatures, thereby alleviating light and temperature limitations and enhancing growth. The 3D model also resolved horizontal heterogeneity in filament distributions, showing accumulation along lake edges where water depths were shallower than the photic zone, enhancing light availability for filaments compared to the central, deeper regions where light limitation was more pronounced. Coupling trait-based 3D IBMs with hydrodynamic models provides valuable insights into the complex interplay between physical and physiological processes that govern cyanobacterial bloom dynamics.