Smart windows are used to reduce energy consumption and improve thermal and visual comfort mainly by controlling the solar flux entering into a building. This article presents a simulation study in which the impact of the applied control strategy on the overall energy consumption (heating, cooling and lighting) is investigated. A commercial building located in Montreal (Canada) with south-oriented integrated electrochromic windows is modeled. The hour-by-hour state of the smart windows required to minimize overall energy consumption while respecting constraints related to thermal and visual comfort is determined through an optimization strategy based on genetic algorithms (GA). Then, this quasi-optimal control is compared to other approaches that could be applied in real-time applications: (i) two types of rule-based controls (RBC), i.e. RBC1 and RBC2 and (ii) a model predictive control (MPC). The impacts of thermal mass and installed light power density are also analyzed. Results show that the four control strategies under study presented similar energy consumption with differences in total energy consumption ranging from 4% to 10%. While more complex controllers such as MPC could potentially lead to improved performances considering more design variables, complex models and extensive commissioning, this study illustrates that simpler control strategies such as RBC2 can also lead to satisfying results.