The future design of high performance buildings is expected to involve more active facade technologies, acting in intelligent collaboration with the HVAC and lighting systems to produce comfortable indoor environments with reduced energy consumption. Integrated control of active facade systems and HVAC is challenging, particularly with thermally-massive HVAC systems such as radiant floors and ceilings. This paper describes methods for devising near-optimal controllers for such integrated systems, allowing for any arbitrary level of complexity in the facade system. An offline-optimization approximation to model predictive control is used with a model consisting of a reduced- order approximation of the zone and HVAC thermal properties and an interpolation grid of the daylight and solar gains attributes of the facade in its various possible states. The optimization over the 24-hour prediction horizon is split into two levels, with GenOpt used at the top level to deal with the complexity of the facade, alongside a linear programming solution to the chilled slab control. The model can be calibrated to match monitored data, or some combination of whole-building energy modeling and Radiance outputs. To test the methods and to estimate energy savings potential, case studies were performed with a calibrated model based on an EnergyPlus ASHRAE 90.1-2010 office building, modied to use radiant slabs and operable Venetian blinds (either internal or external) or electochromic glazing. Results are shown for four US climates. Further research is discussed.