Multi-evaporator vapor compression systems (ME-VCS) are inherently multiinput multi-output (MIMO) systems, often with complex, highly coupled dynamics. Thus, they require more sophisticated control schemes than traditional on-off logic, or decentralized proportional-integral controllers. Unfortunately, many MIMO control design techniques are not well suited for this problem since they require complex numerical computations that do not scale gracefully for the high-dimensional dynamics of ME-VCS systems. This paper exploits the observed similarity of the room dynamics to reduce the computational complexity of designing controllers. We use a linear matrix inequality based controller synthesis technique that exploits symmetry for designing controllers for large-scale ME-VCS systems. This controller synthesis technique was applied to an MEVCS system with 50 rooms. Using tradition control design methods required 41 hours to synthesize a controller, while our technique designed an identical controller in less than 1 second.