This paper addresses optimal design of windows (types of glazing and cavity gas) using BIM, Genetic algorithm and Pareto optimality. For architectural modeling and energy simulation, Revit Architecture and EnergyPlus were selected in this study. In the paper. presented is an automatic file converter, operating in MATLAB GUI (Graphical User Interface) platform, which transforms a gbXML file to an IDF file (EnergyPlus input file). In the paper, a case study is shown to find optimal windows for a sample library building (floor area: $7,235m^2$, 5 stories). The selected performance criteria are energy use and PMV (Predicted Mean Vote), leading to multi-criteria optimization problem. The GA (Genetic Algorithm) was chosen for this minimization problem, while Pareto optimality was used for the multi-criteria problem. It is shown in the paper that optimal Pareto set can be efficiently found and meaningful information is delivered to DM(Decision Maker) using integration of GA, Pareto optimality and BIM.