The purpose of this research was to evaluate the accuracies of building energy simulation programs when modeling glazing units and windows with and without shading devices and daylighting. Program outputs were compared with data from experiments performed at two research facilities designed for these types of studies;Six experiments were performed in a test cell located on the Swiss Federal Laboratories for Material Testing and Research (EMPA) campus in Duebendorf, Switzerland. After determining thermophysical properties, the thermal bridges of the test cell were analyzed using three dimensional heat transfer simulations. The experimental series started with a steady-state and a transient experiment to characterize the cell. A preliminary study was then performed to assess various solar radiation models for tilted surfaces and determine which model performed best. Additional experiments were then run to evaluate solar gain models in building energy simulation programs starting with the simplest case and increasing the complexities of the experiments. These experiments were performed using a solar selective glazing unit without shading, with external and internal diffuse shading screens, and with external Venetian blinds with the blind slats horizontally positioned and tilted 45° downward with the outer slat edge pointed towards the ground. Increasing the complexities of subsequent experiments allowed for careful assessments and diagnoses of the results;An additional experiment was performed at the Iowa Energy Center's Energy Resource Station (ERS) located on the Des Moines Area Community College campus in Ankeny, Iowa USA. In this study, three different windows in combination with three window shading devices were used to evaluate daylighting performance and associated interactions in six exterior test rooms;Robust experiment and sensitivity analyses were used to evaluate the impact of experimental uncertainties for the experiments and program inputs for all of the experiments. A set of comprehensive statistical parameters was used to compare results building energy simulation programs with the experiments and a 95% level of significance was used to determine whether the programs were validated or not. Up to four programs were evaluated in this study for each experiment, including: EnergyPlus, DOE-2.1E, TRNSYS-TUD, and ESP-r.