By coupling algorithmic optimization with parametric modelling and performance simulation, the building design process becomes a series of informed decision-making. Algorithmic optimization helps architects find high-performance design choices which balance complicated trade-offs among various performance objectives. Although algorithmic-optimization aided design is more competent than the traditional manual method, due to complexities in building design, optimization frameworks are not easy to formulate. Focusing on the correlation between energy efficiency, visual comfort, and construction material cost, the current author formulates a bi-phase optimization framework to search for facade fenestration geometries. Afterwards, this optimization framework is applied to one representative office building in China. Optimal design choices are found to have much lower thermal loads and construction cast, but higher visual comfort. This paper highlights how the optimization framework regarding facade fenestration is formulated and how it performs during the building design process.